The physico-chemical nature of the chemical bond:
valence bonding and the path of physico-chemical emergence
by
Martha Lynn Harris
A thesis submitted in conformity with the requirements for the degree of
Doctor of Philosophy
Graduate Department of
the Institute for History and Philosophy of Science and Technology
University of Toronto
© Martha Lynn Harris 2008
The physico-chemical nature of the chemical bond:
valence bonding and the path of physico-chemical emergence
Martha Lynn Harris
Doctor of Philosophy, 2008
Institute for History and Philosophy of Science and Technology
University of Toronto
Abstract
Through the development of physical chemistry and chemical physics over the
late-nineteenth and early-twentieth centuries, the relationship between physics and
chemistry changed to create a broad interdisciplinary framework in which chemists and
physicists could make contributions to problems of common value. It is here argued that
evolving disciplinary factors such as physical and chemical responses to the atomic
hypothesis, the nature of disciplinary formation in Germany and the United States, the
reception of quantum mechanics within physics and chemistry, and the application of
quantum mechanics to the problem of chemical bonding by physicists and chemists,
formed the chemical bond into a physico-chemical theory.
In the late nineteenth-century context of early physical chemistry, the chemical
bond was known as a physical link between atoms, which could not be studied by
chemical means because of the lack of an adequate atomistic framework. Both chemists
and physicists broadly accepted the atomistic hypothesis following the discovery of the
electron at the turn of the twentieth century, which afforded theoretical study of chemical
bonding. Between 1916 and 1919, Gilbert N. Lewis and Irving Langmuir proposed the
valence bond to be a pair of electrons shared between two atoms, within the context of a
ii
cubic model of the atom. However, the lack of a physical mechanism for the shared
electron pair prevented the formation of a fully physico-chemical view of bonding. In
1927, physicists Walter Heitler and Fritz London showed the stability of the valence bond
was caused by the wave mechanical phenomenon of resonance. Chemist Linus Pauling
extended their treatment of the valence bond to a theory of structural chemistry in The
Nature of the Chemical Bond. His synthesis of the physical and chemical views, his value
as a physico-chemical researcher during the 1930s, and the research of his
contemporaries John Slater and Robert Mulliken show that a true physico-chemical blend
was only realized within the amorphous discipline of chemical physics. Finally, it is seen
that this interdisciplinarity of chemical bonding and its supporting framework force a re-
evaluation of the reductionist criteria, and a re-definition of the chemical bond as a
physico-chemical work.
iii
Acknowledgements
I thank my parents, Dr. William Edgar Harris and Dr. Gretchen Luft Hagen
Harris, for the many ways they have supported me during my undergraduate and graduate
education. This thesis would not have been possible without their support and guidance. I
thank my sister, Glenna Catherine Harris, for her support as a roommate, friend, and
fellow doctoral candidate during the writing process. Thank you to my grandfather, Dr.
Walter Edgar Harris, for sharing conversations about his experiences as a chemist,
graduate student, and teacher as I wrote this thesis. Thank you to my aunt, Margaret
Harris, for sharing her interest in the history and philosophy of science and tales of her
graduate studies at the University of Toronto, and for always asking me for the latest in
pop culture.
I thank my committee members Dr. Margaret Morrison and Dr. Craig Fraser for
their support during meetings, for their reading of earlier drafts of this thesis, and for their
instrumental guidance during preparation for my doctoral specialist exam that formed the
basis of this project. Thank you to Dr. Mary Jo Nye, Dr. Anjan Chakravartty and Dr.
Donald Cotter, for their advice and comments on an earlier paper on reductionism and the
chemical bond, which has been incorporated into this thesis.
Thank you to friends and fellow students, Charissa Varma, Victor Boantza, Shana
Worthen, Marionne Cronin, Jenny Crnac, Trevor Buttrum, Gill Gass, Isaac Record, Brigit
Ramsingh, and the students of the IHPST common room for providing many welcome
conversations, distractions and moral support during my time in Toronto.
And finally, I thank my supervisor, Dr. Trevor Levere, for his unfailing support
and guidance during my M.A. and Ph.D. studies, and for teaching me about many things.
iv
Table of Contents List of Figures................................................................................................................... vi Introduction....................................................................................................................... 1 Chapter 1: Physical chemistry and the origins of valency.......................................... 12
1. Physical chemistry and the Ionists. ................................................................................... 12 2. Nineteenth-century debates and the problem of the atom .............................................. 24 3. Growth of a new discipline: the institutional framework................................................ 37
Chapter 2: Atomism, chemical autonomy and the valence bond:.................................. Lewis, Langmuir and the cubic atom............................................................................ 48
1. Introduction......................................................................................................................... 48 2. The cubic atom and early valence bond theory................................................................ 51
2.1 Lewis’ academic influences .......................................................................................................... 51 2.2 Lewis, the atom, and the molecule................................................................................................ 57
3. Acceptance of the cubic atom............................................................................................. 68 3.1 Langmuir and the octet theory of valence..................................................................................... 68 3.2 Priority dispute: Lewis and the reception of the octet theory........................................................ 76
4. Emergence of chemical autonomy ..................................................................................... 79 Chapter 3: Seeds of emergence: wave mechanics meets the valence bond................ 88
1. Introduction......................................................................................................................... 88 2. Wave mechanics meets the chemical bond ....................................................................... 89
2.1 Wave mechanics and the Heitler-London treatment ..................................................................... 89 2.2 The physico-chemical community responds. ................................................................................ 97
3. Pauling’s early career ....................................................................................................... 104 3.1 The ‘boy professor’ meets physical chemistry............................................................................ 104 3.2 Atoms in the grasp: physical chemistry meets the new physics.................................................. 109 3.3 The chemical bond leads to physics: Pauling as theoretical physicist ........................................ 115
4. The physical nature of the chemical bond ...................................................................... 127 4.1 Pauling’s program begins............................................................................................................ 127 4.2 The nature of the chemical bond................................................................................................. 133 4.3 The ‘physico-chemical’ nature of the valence bond ................................................................... 141
Chapter 4: Paths to interdisciplinarity: the chemical bond in context.................... 146 1. Introduction....................................................................................................................... 146 2. Reductionism vs. interdisciplinarity................................................................................ 150 3. Alternate avenues: the chemical bond elsewhere ........................................................... 159
3.1 John Clarke Slater and the valence bond .................................................................................... 159 3.2. Molecular orbitals and Robert Mulliken .................................................................................... 169
4. Chemical bond in demand: Pauling as job candidate.................................................... 176 5. Chemical bond in context: reductionism, interdisciplinary, and the path of disciplinary change..................................................................................................................................... 189
Bibliography .................................................................................................................. 196
v
vi
List of Figures
Page 121: Figure 1: Image of Linus Pauling in doctoral robes, 1925. Courtesy Ava Helen and Linus Pauling Papers, Oregon State University Libraries, Photo 1925i-010-[3806].
Page 127: Figure 2: Image of Fritz London, Ava Helen Pauling and Walter Heitler
on a hiking trip during Ava Helen and Linus Pauling’s visit to Zürich, 1926. Courtesy Ava Helen and Linus Pauling Papers, Oregon State University Libraries, Photo 1926i-116-[3554].
Introduction
The chemical bond, the link formed between two or more atoms that allows them
to be regarded as a molecule, an aggregate of atomic particles with unique chemical
significance, has enjoyed a complicated history. As a scientific object,1 the chemical
bond has been shaped by contributions from physics and chemistry, and in the process
has influenced the course of physico-chemical disciplinary change over the early
twentieth century. At first regarded in the nineteenth century as a phenomenon that was
incapable of study by chemical means, the chemical bond became in the twentieth
century one of the most progressive areas of chemical research.
Accompanying and facilitating this transformation were a number of significant
disciplinary and conceptual changes for chemistry and physics. Spanning the
transformation from an older physical chemistry to a new chemical physics,2 and the
move from quantum theory to quantum mechanics, the chemical bond offers historians
and philosophers of science a valuable opportunity to study the disciplinary issues at
work in the development of the theory itself as well as the disciplinary framework
supporting it. This study is concerned with the nature of the changes in the intermediate,
transitional period from about 1900 to 1935. Here we bring to light a series of contrasts
between physical chemistry and chemical physics, as the disciplinary context for the
historical growth of a physico-chemical theory.
1 For a treatment of how to examine discoveries and theories as scientific objects, see Lorraine Daston, Biographies of Scientific Objects (Chicago: University of Chicago Press, 2000). In particular, see contributing articles by Jed Buchwald, “How the Ether Spawned the Microwordl”, 203-225, and Rivka Feldhay, “Mathematical Entities in Scientific Discourse”, 42-66. 2 This introduction will involve a fuller discussion of the origins and definitions of these disciplines.
1
Physical chemistry was first formed in the 1880s, by a trio of former organic
chemists who named themselves the Ionists: Wilhelm Ostwald, Svante Arrhenius, and
Jacob van’t Hoff. Unhappy with the then current state of chemistry, which had become
dominated by the classification, analysis and synthesis of organic substances, Ostwald
and his men envisioned a new, theoretical approach to chemistry where physical
principles could be applied to chemical problems. The Ionists, so named because of their
intense study of the properties of ions in solution, hoped to gain a foundational
understanding of the reaction process and the state in which reactions occur. In this way,
John Servos has argued, the early physical chemists sought to redirect their science back
towards an older, Laplacian tradition of the nature of chemical affinity, to look at the
mechanism of chemical change rather than typify and classify the products.3
In a different kind of manifestation of the relationship between physics and
chemistry, chemical physics first developed broadly in the 1930s, driven by the response
by physicists, chemists and mathematicians who responded to new research questions of
collaborative value for their fields. This new growth was influenced by the discovery of
new problems, particularly those arising from quantum mechanics, and the rise of
physical techniques like molecular spectroscopy and x-ray crystallography. The
molecular orbital and valence bond theories of the chemical bond developed along with
the discipline, and formed an important part of the first published chemical physics
literature in the early 1930s, when they were robustly developed by Erich Hückel and
3John W. Servos, Physical Chemistry from Ostwald to Pauling: The Making of a Science in America (Princeton University Press, 1990).
2
Robert S. Mulliken (molecular orbital), and Linus Pauling and John C. Slater (valence
bond). 4
The most complex issue when discussing the formation of physical chemistry and
chemical physics is how to determine their respective identities as disciplines. In her
interpretation of physico-chemical disciplines, Mary Jo Nye suggests that historical
discipline formation can be characterized by any or all of six criteria: the presence of a
historical genealogy, a core literature, common practices and rituals, a physical
homeland, external recognition, and shared values. 5 Nye’s approach offers enough
complexity to treat broad fields of scientific study such as physical chemistry and
chemical physics, but also the flexibility to differentiate between them. The histories of
these two sciences both exhibit components from this scheme, but the different ways the
components manifest themselves show that physical chemistry and chemical physics
have very distinct identities.
While chemical physics lacked, as a whole, the laboratory homeland that defined
physical chemistry so well in its formative years, it had a well-defined point of origin in
the birth of the Journal of Chemical Physics in 1933. Physical chemistry also had its
defining journals, most notably the Zeitschrift für physikalishe Chemie, founded in 1887,
can be defined very amorphously in concept, as Keith Laider shows, set apart from Nye’s
scheme, since chemists and physicists themselves had differing opinions on the subject
over the course of the late 19th and 20th centuries. Even Gilbert Lewis, a main protagonist
4 Chapter 4 contains a more extensive look at the contrasting nature of these two theories of chemical bonding, specifically in the differing approaches taken by Mulliken, Pauling and Slater. 5 Mary Jo Nye, From Chemical Philosophy to Theoretical Chemistry: Dynamics of Matter and Dynamics of Disciplines (1800-1950) (University of California Press, 1993)
3
in our chemical bond story, once defined physical chemistry as simply “everything that is
interesting.”6
When contrasting physical chemistry and chemical physics, it is clear from
studies by Servos, Nye, Laidler, Colin Russell and Alan Rocke that although both
disciplines were motivated by combining chemical and physical work, the end results had
different levels of success outside of chemistry proper.7 Physical chemistry was created
by chemists, and their motivation was largely internal to their field: dissatisfaction with
the dominance of organic studies. While the external influences of physics, such as
theoretical tools and experimental methods, helped them create a new form of chemistry,
a form of chemistry was all it remained. Broader disciplinary cross-pollination was never
accomplished and physical chemistry made small impact in departments of physics. The
men who contributed to chemical physics came from backgrounds where connections
between chemistry and physics were already known, especially in the context of the
quantum theory. Their initial identities as chemists, physicists, or physical chemists all
allowed them opportunities to partake in interdisciplinary research.
Even when defined in a more general way as theoretical chemistry, Nye’s study
shows that Ostwald’s work was only one part of a movement towards more theoretical
research over the late nineteenth century. Her examples of the research schools of
Christopher Ingold (London), Arthur Lapworth (Manchester), and Robert Lespieau
(Paris) matured in the early twentieth century within a previously rigid boundary between
physical and organic chemistry. These groups used physical methods to study classical
6 Keith J. Laidler, The World of Physical Chemistry (Oxford: Oxford University Press, 1995), 7. 7 Colin A. Russell, The History of Valency (Leicester University Press, 1971), Alan Rocke, Chemical Atomism in the Nineteenth Century: From Dalton to Cannizzaro (Ohio State University Press: Columbus, 1984).
4
organic compounds and reactions, and Ingold in particular wished to revolutionize
chemical language by studying the importance of the dynamic electron in reaction
pathways. They were a contrast to both the Ionists’ physical chemistry and the quantum-
mechanically rich chemical physics, in developing physical organic chemistry out of the
previous organic tradition.
The nature of chemical physics’ identity struggle, then, is due to its place in what
Buhm Soon Park calls the “borderland” between physics and chemistry, an area that
physical chemistry did not ever fully inhabit.8 Kostas Gavroglu and Ana Simões
characterize chemical physics as a “confluence of diverging traditions”, several fields
blending according to research problems in chemistry, physics and mathematics, rather
than a single group of chemists reaching out for physics.9 Many of the research problems
that contributed to the growth of chemical physics had their seeds in the 1910s and 1920s,
such as x-ray crystallography and quantum theory. These eventually assisted the research
that made up early chemical physics.
Academic institutions play an important role in disciplinary emergence, in the
cases of both physical chemistry and chemical physics. The nature of the path of
institutional change also highlights the contrast between the two disciplines, as we see
physical chemistry relying heavily on a clear laboratory tradition, modeled on the pattern
formed by the original institution for physical chemistry, the University of Leipzig.
Chemical physics, however, spread through a less structured path. The amorphous
character of chemical physics in quantum chemistry will be the focus of much later
8 Buhm Soon Park, Computations and Interpretations: The Growth of Quantum Chemistry, 1927-1967 (Johns Hopkins University: Ph.D. Thesis, 1999). 9 Kostas Gavroglu, and Ana Simões, “The Americans, the Germans, and the beginnings of quantum chemistry: the confluence of diverging traditions”, Historical Studies in the Physical Sciences, 25(1994), 47-110.
5
discussion in chapter 3 and 4, especially in the context of American science in the early
twentieth century. Section 1.4 concentrates on the early spread of physical chemistry and
Ostwald’s ideology, and the work of first-generation Ionist Arthur Noyes in bringing the
Leipzig model to North American laboratories.
The history of the chemical bond, then, speaks to these concerns of identifying
and distinguishing chemistry, physics, chemical physics and physical chemistry, and the
conceptual shifts experienced by physicists of this period are as crucial as those in
chemistry in understanding why the chemical bond developed in the way it did. Works by
Mara Beller, Olivier Darrigol and Helge Kragh show that the transformation from
quantum theory to quantum mechanics became the consuming problem for physicists in
the 1920s, particularly for Europeans.10
In this era, physicists struggled with the conceptual shifts precipitated by the
discovery of the electron, Planck’s quantum of action, radioactivity, and finally the
uncertainty principle, as the nineteenth-century view of the world as a mechanical system
and the classical concept of causality were abandoned. Bohr’s orchestration of the
principle of complementarity and his acceptance by those who worshipped him as a hero
promoted an uncritical acceptance of the Copenhagen interpretation. This acceptance,
however, masked the deeper philosophical troubles of quantum physicists trying to find
meaning in the anti-realist, acausal interpretations of the natural world that they were
presented with.11 In much of this history, however, apart from its relevance to atomic
models adopted by members of the chemistry community, it is difficult to isolate a single
10 Mara Beller, Quantum Dialogue: The Making of a Revolution (Chicago: University of Chicago Press 1999); Oliver Darrigol, From c-numbers to q-numbers : the classical analogy in the history of quantum theory (Berkeley: University of California Press, 1992); Helge Kragh, Quantum generations: a history of physics in the twentieth century (Princeton: Princeton University Press, 1999). 11 Beller, Quantum Dialogue.
6
narrative thread that is relevant for chemistry. Until 1927, physical attention was largely
turned toward inward disciplinary reflection; chemistry was at best limited to the
periphery.
The earliest part of the story of the chemical bond begins in the late nineteenth
century, during a time when chemists and physicists began to consider the value of the
atomic hypothesis in their fields. Examining the atom itself, and by consequence the links
between atoms, lay beyond the scope of experimental investigation in the 19th century,
and positivist chemists and physicists rejected the use of atoms in teaching and as a
causal explanation for natural phenomena. Chemists, restricted to knowledge of reactions
with tentative understanding of the reaction process, made few claims about the nature of
the process of bonding behind it. Valency was understood as the grouping of elements
according to their combining power, or affinity, but anything more precise was outside
the limits of chemical investigation. As chemists embraced the atom in the 20th century, it
was most valuable for its ability to explain chemical bonding, as in the work of physical
chemists Gilbert N. Lewis and Irving Langmuir. The value of an explanation for bonding
set the chemists’ perspective on the atom apart from the physicists, during the first few
decades of the twentieth century: physicists sought to discover the whole internal
structure of the atom, while chemists needed to know only enough to understand the
process of molecule formation.
The point where theoretical physics enters our chemical bond story, in 1927, has
been most sensational from a philosophical rather than historical perspective. 1927 was
the year when German physicists Walter Heitler and Fritz London applied Erwin
Schrödinger’s wave mechanics to the single, non-polar bond in the hydrogen molecule.
7
Their treatment was essentially the formal application of a physical theory to a chemical
problem. Historically regarded as the seed of quantum chemistry, a sub-discipline of
chemical physics, Heitler and London’s work has also been regarded as the basis of the
argument that chemical theory may be reduced to physics. Philosophers and some
historians of chemistry have confronted this view in recent years.12 They suggest
reductionism cannot be achieved in practice due to difficulties in solving the Schrödinger
equation, and that there are differences in chemical practice and culture that cannot be
captured by a physical theory. 13,14 These authors illustrate how valuable the chemical
bond is for the philosophy of chemistry and exhibit alternative ways of understanding the
relationship between the two disciplines. Anti-reductionism forms a major part of this
study, in interpreting the relevance of Heitler and London’s discovery for quantum
chemistry.
One of the best approaches to analyzing the chemical bond in the context of
chemical physics is Nye’s focus on the scientific tradition within a disciplinary
framework. The research schools in her study, and that of the Ionists, passed down a
tradition of scientific work through experimental practice and a disciplinary ideology. In
a similar vein, Gavroglu further suggests that the way chemistry differentiates itself from
12 See Nikos Psarros, “The Lame and the Blind, or How Much Physics Does Chemistry Need?”, Foundations of Chemistry, 3 (2001), 241-249; Eric. R. Scerri, “Has Chemistry at Least Been Approximately Reduced to Quantum Mechanics?” PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association, Vol. 1994 (1994), 160-170; Ana Simões, and Kostas Gavroglu, “Issues in the History of Theoretical and Quantum Chemistry, 1927-1960”, In Carsten Reinhardt (ed.), Chemical Sciences in the 20th Century: Bridging Boundaries, (Wiley-VCH, 2001); Andrea. I. Woody, “Putting Quantum Mechanics to Work in Chemistry: The Power of Diagrammatic Representation”, Philosophy of Science, 67 (2000) (Proceedings), S612-S627. 13 Woody, “Putting Quantum Mechanics to Work in Chemistry”; Scerri, “Has Chemistry at Least Been Approximately Reduced to Quantum Mechanics?”. 14 Psarros, “The Lame and the Blind”; Kostas Gavroglu, “The Physicists’ Electron and its Appropriation by the Chemists”, in Histories of the Electron, Jed Buchwald and Andrew Warwick (editors) (Cambridge: M.I.T. Press, 2001), 363-400.
8
physics through its community and culture forms a barrier to the reductionist view.15
While chemists and physicists both rely on entities like the atom and electron, Gavroglu
believes chemists have essentially appropriated them from physics because they help
explain a whole class of chemical problems, namely those of chemical bonding.
Within chemical physics, these interdisciplinary studies themselves fell into
separate traditions of the molecular orbital and valence bond theories.16 While each were
accompanied by different experimental methods in the early stages (molecular
spectroscopy and x-ray crystallography, respectively), the preference for one theory over
another in a divided chemical and physical community was largely motivated by the
predictive value of each theory, rather than an allegiance to a previous experimental
tradition.17 While the physical homeland of Leipzig was a recognizable nexus for the
Ionists’ tradition of physical chemistry, the chemical bond, with its predominant
theoretical and physical influences, grew into a physico-chemical framework far different
from what Ostwald created. From these historical studies comes the question of whether
it is possible for a theoretical tradition to emerge within this disciplinary framework.
Broadly, the chemical bond provides an opportunity to study the disciplinary
relationship between chemistry and physics. From the historical perspective, the chemical
bond is seen as the fruit of interdisciplinary growth in the early twentieth century, but
from the philosophical perspective it has been seen by some authors as grounds for the
15 Gavroglu, “The Physicists’ Electron.” 16 Both molecular orbital and valence bond theories used the wave mechanical properties of electrons to build a model of a bonded molecule. However, in the valence bond conception the molecule is seen as a sum of individual, directed bonds between atoms, rather than as a holistic entity composed from the top-down as in the molecular orbital method. Molecular orbitals in the work of Robert S. Mulliken, in the context of the contrast with the valence bond in the work of John C. Slater and Linus Pauling will be studied in more detail in chapter 4. 17 Sason Shaik, and Philippe C. Hiberty, “Valence Bond Theory, Its History, Fundamentals, and Applications: A Primer”, Reviews in Computational Chemistry, 20 (2004), 1-119; Gavroglu and Simões, “Issues in Theoretical Chemistry”.
9
reductionist belief that, at its basic level, chemistry is purely physics. As well, the
contrasting character of physical chemistry and chemical physics suggests the eventual
maturation of the chemical bond within chemical physics, not physical chemistry, was
influenced by historical factors not in place at the time of early physical chemistry. The
differing cultural practices between physics and chemistry, as argued by Nye, Woody,
Psarros, and Gavroglu, raise the question of disciplinary ownership of scientific work. In
the case of the chemical bond, is it possible for one discipline to claim ownership of a
theory? How do traditions form around works of theory, rather than works of
experiment?
This study will seek to answer these questions through an examination of the
history of the valence bond during the early twentieth century. The valence bond was first
known in chemistry, as a chemical entity, about which knowledge was limited. The
factors that influenced its transformation into a physico-chemical entity remain to be fully
understood. Here, it will be seen that the interdisciplinary nature of valence bond research
takes on two forms: first, the character of the valence bond itself, as given by Lewis,
Heitler, London, and Pauling, as a physico-chemical theory, and second, the physico-
chemical character of the two disciplines in which it developed, physical chemistry and
chemical physics. It is the former that will be of primary concern. This story is less one of
interdisciplinary identity but of the genesis of an interdisciplinary theory.
Chapter 1 will examine the disciplinary context of early physical chemistry, and
the late nineteenth-century background to the chemical bond, of valency and atomism. In
this period we see the formation of a disciplinary framework for rigorous physico-
chemical work, that lacked an atomistic framework adequate for the explanation of
10
valence bonding. Chapter 2 looks at the discovery of the valence bond proper, in the
work of Gilbert Lewis and Irving Langmuir. While this period saw the widespread
acceptance of the atomic theory in physics and chemistry, and the rise of an autonomous
chemical framework of atomism, the lack of a physical explanation for the mechanism of
bonding prevented the realization of a fully physico-chemical worldview. This came
about during the events of Chapter 3, as Linus Pauling extended Heitler and London’s
wave mechanical treatment of the valence bond to a theory of structural chemistry in The
Nature of the Chemical Bond. His synthesis of the physical and chemical views and his
value as a physico-chemical researcher during the emergence of chemical physics
indicates a true physico-chemical blend was realized during the early 1930s. Finally, in
Chapter 4, it is seen that these issues of interdisciplinarity force a re-evaluation of the
reductionist criteria, and a re-definition of the chemical bond as a physico-chemical work.
The history of the development of the valence bond raises historical and
philosophical questions about the definition of interdisciplinary research. Here, over the
course of three decades, we see chemists and physicists working with physical methods
to study chemical problems. The example of the valence bond through Lewis, Heitler,
London and Pauling exemplifies one tradition of theoretical research. What makes their
chemical bond physico-chemical is the presentation of concepts and values from both
physical and chemical traditions. More than a source of theory reduction, and more than
the simple application of physical methods to chemical means, the chemical bond is a
theory at the centre of a host of issues common to chemistry and physics. Taking into
consideration its disciplinary origins and development, we see that the chemical bond is
neither physical, nor chemical, but a truly physico-chemical theory.
11
Chapter 1: Physical chemistry and the origins of valency
“[C]hemistry – the science of the changes which bodies undergo when being combined or separated – deals with the most complicated side of reality.”18
1. Physical chemistry and the Ionists.
The first concerted attempt to unify physics and chemistry in a formal disciplinary
sense came in the form of physical chemistry.19 This discipline was born in 1887, when
Wilhelm Ostwald and Jacobus van’t Hoff founded the first journal for physico-chemical
studies, the Zeitschrift für physkalische Chemie, and the same year when Ostwald opened
the first laboratory devoted to the study of physical chemistry. At this time Ostwald, van’t
Hoff, and their contemporary Svante Arrhenius were young chemists who had settled into
their early careers, but had grown dissatisfied with the state of their discipline. In the
previous decades, chemistry had become dominated by organic studies, but instead of
cataloguing species like botanists in a chemical laboratory, Ostwald, Arrhenius and van’t
Hoff began a movement to study the principles by which those species were formed,
under the framework of a new, theoretical, foundational, allegemeine Chemie. Having
stated in his Master’s examination a decade earlier, “modern chemistry is in need of
reform”,20 Ostwald led the group in reforming their science by borrowing the principles
from another: physics.
18 John Thedore Merz, A History of European Scientific Thought in the Nineteenth Century, Volume 1 (Dover: New York, 1965) (2 vol reprint), on 390. 19 Previous chemists such as Michael Faraday had considered themselves as part of a broader tradition of natural philosophy, in which there were no sharp divisions between chemistry and physics. In Faraday’s view, concepts like affinity were not strictly chemical or physical but related to underlying natural forces that were universal. Levere, Trevor, Affinity and Matter: Elements of Chemical Philosophy 1800-1865 (Oxford: Clarendon Press, 1971), 69-72. 20 Servos, Physical Chemistry from Ostwald to Pauling, 3.
12
For Ostwald, this came from an early interest in chemical affinity, while as a
student at the University of Dorpat in the 1870s. Despite a general awareness of chemical
affinity as a cause of chemical action and some early researches about the concept,
affinity remained a largely undefined phenomenon.21,22 Few chemists “had full
command” of the research that had accumulated, “and no one had developed a conceptual
framework into which the evidence could be incorporated.”23 By the time Ostwald
approached it, there was evidence to suggest affinity was correlated with heat, energy,
and electricity. This led him to thermodynamics, and what J.R. Partington has called
“[o]ne of the most remarkable periods in the history of physical chemistry.”24
The first major studies of thermodynamics were made in 1824 by French physicist
Sadi Carnot, as the study of the interconversion between heat and work, particularly as
seen in heat engines.25 Thermodynamics had a broader impact in physics much later in
the century, following research by English physicists William Thomson (Lord Kelvin)
and James Joule in the 1850s. Joule proposed that heat was a form of motion, which
Kelvin used to re-interpret Carnot’s original studies. Kelvin discovered in this process
that when heat flows from a warmer to a colder body there is no work done and no loss of
energy, but instead a “loss in the capability to do work”, which German physicist
21 For a comprehensive account of the concept of affinity preceding the chemical revolution and through the work of Davy, Berzelius and thermochemistry, see Michelle Goupil, Du Flou au Clair (Editions du comité des travaux historiques et scientifiques: Paris, 1991). A study of chemical affinity in the 17th and 18th centuries, through the work of Lavioiser, can be found in Mi Gyung Kim, Affinity, That Elusive Dream: A Genealogy of the Chemical Revolution (Cambridge: M.I.T. Press, 2003). For a history of affinity in the 19th century through Davy, Faraday, Berzelius, Whewell and Berthollet, see Trevor Levere, Affinity and Matter: Elements of Chemical Philosophy 1800-1865 (Oxford: Clarendon Press, 1971). 22 See Helge Kragh, “Van’t Hoff and the Transition from Thermochemistry to Chemical Thermodynamics”, in W. Hornix and S.H.W.M. Mannaerts (editors), Van’t Hoff and the Emergence of Chemical Thermodynamics: Centennial of the first Nobel Prize for Chemistry 1901-2001 (Delft: D.U.P. Science, 2001), 191-211; Laidler, The World of Physical Chemistry; Nye, From Chemical Philosophy to Theoretical Chemistry. 23 Servos, Physical Chemistry from Ostwald to Pauling, 18. 24 J.R. Partington, A History of Chemistry, Volume 4 (London: MacMillan & Co. Ltd, 1964), 640. 25 Laidler, The World of Physical Chemistry, 83.
13
Rudolph Clausius defined in 1854 as entropy. 26 The work of these three men formed the
modern understanding of the first two laws of thermodynamics, which Clausius stated
succinctly as: “[t]he energy of the world is constant; the entropy tends toward a
maximum.”27
Much of modern chemical thermodynamics originates in the work of American
theoretical physicist J. Willard Gibbs, who wrote a series of papers in the 1870s defining
the thermodynamic processes at work in chemical reactions.28 These papers were largely
formal mathematical works in which Gibbs derived through calculus a general
understanding of the essential laws of chemical equilibrium, and defined the physical
processes at work in a reaction in terms of differential equations, the properties of
integrals, and geometrical forms.29 All of these involved the five variables volume,
pressure, temperature, energy (ε) and entropy (η): Gibbs called these “state functions”
because their values were independent of the path taken by the system, or chemical
reaction. Work and heat, by contrast were “determined by the whole series of states
through which the body is supposed to pass.”30 This led Gibbs to distinguish them
mathematically as two-dimensional areas, obtained through integration over certain
boundary lines of volume, pressure, or temperature. In this way he constructed the
26 Laidler, The World of Physical Chemistry, 102. 27 As quoted in J. Willard Gibbs, Scientific Papers (Dover Publications: New York, 1961), 55: “Die Energie der Welt ist constant; die Entropy strebt einem maximum zu.” 28 Reprinted together in Gibbs, Scientific Papers. 29 For further studies of Gibbs’s papers and their impact on the Ionists, see Alexander Y. Kipnis, “Early Chemical Thermodynamics: Its Duality Embodied in Van’t Hoff and Gibbs”, in W. Hornix and S.H.W.M. Mannaerts (editors), Van’t Hoff and the Emergence of Chemical Thermodynamics: Centennial of the first Nobel Prize for Chemistry 1901-2001 (Delft: D.U.P. Science, 2001), 212-242., and R.J. Deltete, “Josiah Willard Gibbs and Wilhelm Ostwald: A contrast in scientific style”, Journal of Chemical Education 73(1996), 289-294. 30 Gibbs, Selected Papers, 3.
14
relationships that Joule, Kelvin and Clausius discovered, in a new, mathematically
defined chemical landscape.
Conceptually, one of the most important aspects of Gibbs’s work was his
treatment of the environment in which the reaction took place. This came through in his
construction of a mathematical object called a thermodynamic surface, which represented
a system in the state of thermodynamic equilibrium. For a system at equilibrium between
two physical states (e.g. solid and liquid), the surface is a line, but for three states (solid,
liquid, gas) it is a plane triangle. In his second paper he studied the properties of such a
surface in “a medium of constant temperature and pressure.” Using the properties of the
surface and its careful definition in terms of state functions, Gibbs went on in his third
paper to study the criteria needed for a system to reach thermodynamic equilibrium, and
the “special laws which apply to different classes of phenomena.”31 Gibbs’s innovative
study of the reaction environment was technically challenging, but conceptually
influential on the Ionists. With chemical thermodynamics, chemical reactions could be
studied through the physical variables of the reaction state, rather than the chemical
definitions of the reactants.
The science of thermodynamics was a way of studying a physical or chemical
system in terms of the bulk properties of matter, such as temperature, pressure, and
energy. It was very theoretical and mathematical, and required a conception of reactions
in the abstract: quantities measured in the laboratory would be interpreted in terms of new
fundamentals like entropy and work, which were related by mathematical analysis to
more recognizable variables like volume and temperature. This lent a quality of
intangibility that made it difficult to interpret. Too many chemists “felt no need yet to go 31 Gibbs, Selected Papers, 61-2.
15
beyond their usual subjects of research” in order to learn it, and to the average chemist,
who had very little understanding of calculus, Gibbs’ work was incomprehensible.32 It
was decades before the discipline of chemistry as a whole was able to understand it, as it
“required considerable mathematical knowledge to follow his reasoning”, and Gibbs
consequently “had no impact on physico-chemical reasoning.”33
The place where chemical thermodynamics had the most impact was in the work
of the Ionists, who applied it to aspects of chemistry that had not previously been well
understood, such as chemical affinity. Danish chemist Julien Thomsen and French
chemist Marcellin Berthelot had recently approached affinity from the perspective of
what became known as thermochemistry. Their treatment of affinity was based on the
law of energy conservation and what became known as the Thomsen-Berthelot principle,
which stated that all chemical reactions were exothermic, or accompanied by the
production of heat. In 1865 Berthelot formalized this in the principle of maximum work,
that every chemical change accomplished without the intervention of external energy
tends toward the process that disengages the most heat. Through the 1860s and 1870s
chemical affinity was assumed to operate within these thermal constraints, as an “innate
attractive force”, where the heat of reaction was “the result of a transformation from
potential to kinetic energy.”34 At about the same time, in 1864, Norwegians Cato
Guldberg and Peter Waage had discovered through a study of reaction rates that the rate
of reaction was correlated to the activity, or level of dissociation of molecules in solution.
Guldberg and Waage’s result, known as the law of mass action, spoke to just the
32 Kipnis, “Early Chemical Thermodynamics”, 219. 33 Kragh, “Van’t Hoff and the Transition from Thermochemistry”, 194. 34 Kragh, “Van’t Hoff and the Transition from Thermochemistry”, 192-3.
16
questions being raised about the reaction state in thermodynamics, but their work
received almost no attention until the Ionists made their own discoveries in the 1880s.
When Ostwald first approached affinity in the 1870s, his interest was in finding
quantitative values that could be associated with the qualitative changes that
accompanied chemical reactions. These could be measured in terms of physical
properties such as specific volume, refractive index, or change in temperature.35 He
began this line of research in his Master’s thesis by studying a series of acid-base
reactions, to determine their affinity for neutralization: different acids will produce
different amounts of heat when neutralized with the same alkali. By measuring this
change in temperature Ostwald could quantify the affinity between certain acid-base
pairs, and found that the affinity between an acid and base does not depend on the nature
of the base in reaction.36 He made further discoveries about the process of catalysis,
when he found a relationship between the electrical conductivities of weak acids and the
rates of reactions they catalyse. Attaching specific, numerical values to the idea of
chemical affinity, “which had long been referred to in the literature in a qualitative and
often arbitrary way”, redefined an aspect of chemical theory by appealing to physical
properties of the reaction state.37
Ostwald continued this program into the early 1880s, but worked largely
independently from the major circles of German chemistry, largely because he pursued
research that did not fit with that of his contemporaries. However, in the next few years
he made connections with other chemists who were interested in similar questions. The
35 Erwin N. Hiebert, and Hans-Gunther Körper, “Wilhelm Friedrich Ostwald”, Dictionary of Scientific Biography (New York: Scribner, 1981, c1980-c1990), Vol 15, Supplement I, 455-469, on 456-8. 36 Servos, Physical Chemistry from Ostwald to Pauling, 22; Laidler, The World of Physical Chemistry, 279. 37 Hiebert and Körper, “Wilhelm Friedrich Ostwald”, 458.
17
first were Swede Svante Arrhenius, who had begun his career with doctoral work in
physics at Uppsala University, and Dutch chemist Jacobus Henricus van’t Hoff.38
In his dissertation, completed in 1884, Arrhenius studied the electrolytic theory of
dissociation, which stated that the electric current conducted by a liquid solution, such as
in a galvanic cell, was carried by the “active” parts of a solution, now known as ions, in
1887.39 Arrhenius proposed that these ions were in fact always present in a solution.40 He
defined the “active” molecules, which contributed to the electric current, as “those who
have independent motion,” and “inactive” ones as those “whose ions are [still] firmly
bound together.”41 This assumed that some portion of the electrolyte in solution was
dissociated into ions which move about freely in the solution. From 1884 to 1887
Arrhenius studied the level of “activity” of ions in solution, demonstrating that properties
of chemical affinity, reactivity and electrical conductivity could all be understood in
terms of the ions present in solution, and that full ionic dissociation was reached at the
point of extreme dilution of the solute. The ability to determine the concentration and
level of dilution also provided a way of understanding osmotic pressure (the pressure
exerted on a semipermeable membrane dividing two solutions of different concentration)
in terms of the “activity” of the reactants.
It was challenging for Arrhenius to find a receptive audience for his work in
Sweden. His thesis was not clearly written, he was not in good favour with his professors
at Uppsala, and the extent that his dissertation “partook of both physics and chemistry 38 A comprehensive account of the Ionists early careers, their research styles, and the founding of physical chemistry can be found in Robert S. Root-Bernstein, The Ionists: Founding Physical Chemistry, 1872-1890 (Ph.D. Thesis, Princeton University, 1980). 39 Elisabeth Crawford, “Arrhenius, the Atomic Hypothesis, and the 1908 Nobel Prizes in Physics and Chemistry,” Isis 75(1984), 503-522, on 505. 40 Laidler, The Story of Physical Chemistry, 211. 41 H.A.M. Snelders, “Svante August Arrhenius”, Dictionary of Scientific Biography (New York: Scribner, 1981, c1980-c1990) Vol 1, 296-302, on 297.
18
made its acceptance by Swedish physicists and chemists problematic.”42 His
reinterpretation of electrolytic dissociation in terms of ions did not sit well with his
committee, who were more comfortable with the traditional belief that the nature of a
chemical compound remained unchanged even in solution. The idea that potassium and
chlorine, for example, should not be joined together, but exist as different ions in a
potassium chloride solution, to them was nonsense. For this work Arrhenius received a
grade of “approval without praise”, which was not enough to earn him a docentship at
Uppsala.43
Resistance to his work in Sweden forced Arrhenius to look for support elsewhere,
and he quickly found it from Ostwald, who was working energetically from Riga to make
contacts in Europe with like-minded researchers who could bolster support for his new
“allegemeine” chemistry. Ostwald had read Arrhenius’ dissertation, was impressed with
it, and visited the young physicist in Sweden after meeting with others in Denmark and
Norway in the summer of 1884.44 Later that year Arrhenius joined Ostwald at the fall
meeting of the German Naturforscherversammlung in Madeburg, where Ostwald spoke
on his work on chemical affinity, and Guldberg and Waage and Arrhenius’ studies on
electrolytic dissociation. “As a result of this meeting”, Robert Root-Bernstein argues,
“Arrhenius’ research was generally known” in Germany, and through Ostwald he was
offered a docentship in Riga.45 Now satisfied of the value of Arrhenius’ research, the
board at Uppsala University appointed him a lectureship in physical chemistry in 1884. A
42 Laidler, The Story of Physical Chemistry, 217; Crawford, “Arrhenius, the Atomic Hypothesis, and the 1908 Nobel Prizes”, 505. 43 Snelders, “Svante August Arrhenius”, 297; Crawford, “Arrhenius, the Atomic Hypothesis, and the 1908 Nobel Prizes”, 505. 44 Root-Bernstein, The Ionists, 172-3. 45 Root-Bernstein, The Ionists, 173.
19
travel grant in 1885 allowed him further meetings with Ostwald, and opportunities to
develop the electrolytic theory more comprehensively.
As Ostwald and Arrhenius began to collaborate and build a network of contacts in
Europe, they grew to form a trio with the concurrent research of Dutch chemist Jacobus
Henricus van’t Hoff. When van’t Hoff began studying chemical thermodynamics he had
already completed a dissertation in stereo-chemistry, and discovered independently but
contemporaneously with French chemist Joseph Le Bel in 1874 that the optical activity of
carbon compounds was due to the three-dimensional, tetrahedral structure of the carbon
atom. Interested in affinity, he subsequently tried unsuccessfully to quantify the forces
between atoms by studying the structure of organic compounds.46 It was then that he
moved to a thermodynamic approach, influenced by the previous research of Clausius
and Gibbs.47 His major work of 1884, Études de Dynamique chimique (which Laidler
called “the most original book ever written on any aspect of chemistry”48), was a
comprehensive study of the dynamics of chemical reactions. In the Études van’t Hoff
showed that chemical reactions were a dynamic process and, like Arrhenius and Ostwald,
discovered how to define aspects of this process in physical terms of thermodynamics.
Van’t Hoff, like other chemists of the time, “felt alienated” by Gibbs’ “abstract
and mathematically terrifying thermodynamics” and wanted to obtain a clearer
expression of chemical equilibrium.49 While Guldberg and Waage had obtained their law
of mass action by examining physical influences on affinity, van’t Hoff approached the
46 Servos, Physical Chemistry from Ostwald to Pauling, 25-6. 47 Kragh, “Van’t Hoff and the Transition from Thermochemistry”. 48 Keith J. Laidler, “Van’t Hoff and the Birth of Chemical Dynamics”, in W. Hornix and S.H.W.M. Mannaerts (editors), Van’t Hoff and the Emergence of Chemical Thermodynamics: Centennial of the first Nobel Prize for Chemistry 1901-2001 (Delft: D.U.P. Science, 2001), 243-255, on 243. 49 Kragh, “Van’t Hoff and the Transition from Thermochemistry”, 199.
20
same question by looking at the rate of reaction. In 1884 he stated that for an equilibrium
reaction aA + bB ↔ yY + zZ, where the equilibrium constant K is given by
[Y]y[Z]z/[A]a[B]b, the same reaction may be expressed thermodynamically as
d/dt (ln K) = q/RT2,
where R is the ideal gas constant and T is the temperature at equilibrium.
The second major result in the Études was striking a relationship between reaction
rate and concentration as a simple differential equation,
- dC/dt = kC,
which Ostwald called in 1887 the “order of the reaction”: in a first-order reaction the rate is
proportional to the concentration of the reactant, in a second-order theory it is proportional
to the square of the reactant concentration, and so on.50 Lastly, van’t Hoff struck a
connection between ideal gases and dilute solutions. His solution,
PV = iRt,
where V is the volume of a solute instead of a gas, and i is a correlation factor determined
empirically for individual reactions. While the correlation factor meant van’t Hoff could
not obtain an exact equivalence between the properties of ideal gases and solutions, he
showed that the two chemical states were governed by the same physical laws and
constants.
The collective works of Ostwald, Arrhenius and van’t Hoff formed a broad re-
interpretation of chemical processes in terms of their physical properties, and relied on an
understanding of these properties as manifestations of the behaviour of charged ions in
solution. Taken in isolation, the value of their discoveries is not easily evident, because
they brought together several small concepts that had been examined in smaller pieces. 50 Laidler, “Van’t Hoff and the Birth of Chemical Dynamics”, 244.
21
When combined they formed, in Arrhenius’ words, “an imposing and harmonic scheme.”51
Even in the mid-1880s, the three men did not realize the value of their respective works
until they began to correspond with each other, and collaborated to reveal the connections
between the concepts they had developed.
The meaning of the catalytic relationships that Ostwald had discovered in the early
1880s could be explained in terms of the mass action of ions, following his reading of
Arrhenius’ dissertation in 1884, when his electrolytic dissociation theory showed how
reactivity was influenced by the level of dissociation. And though Arrhenius had
understood in 1884 that there was a relationship between the level of dissociation and
osmotic pressure, it was the discovery of van’t Hoff’s expression for the ideal solution that
made him realize there was a key connection between his dissociation law and the
properties of ideal gases (and hence, ideal solutions). Arrhenius proposed an activity
coefficient, α, denoting the amount of dissociation into “active” molecules: upon
discovering van’t Hoff’s work he determined mathematically that this could be used to find
van’t Hoff’s correlation factor i, through the relationship i = 1 + (k-1)α (where k = the
number of ions each active molecule dissociates into). In 1889 Arrhenius further developed
van’t Hoff’s work on reaction rates by showing that the rate of reaction will vary with
temperature, because the equilibrium between normal and active molecules shifts with
temperature.52
Just as their scientific research benefited from their correspondence and
collaboration, so did the standing of the Ionists’ research in the scientific community. None
of the three men hailed from the mainstream of their fields, and were, as Root-Bernstein
51 Partington, A History of Chemistry, 658. 52 Laidler, “Van’t Hoff and the Birth of Chemical Dynamics”, 246.
22
has characterized them, “largely autodidacts and eclectics”, “educated in scientific
backwaters” of their homelands.53 Beginning in the summer of 1884, when Ostwald and
Arrhenius made contact, they gradually built contacts with others of like mind, such as
Guldberg and Waage, who likewise tended to work in the periphery of their own
communities. As the Ionists reciprocated these professional connections, outsiders began to
see evidence that this peripheral research in fact had value within a new, growing scientific
community – hence, Arrhenius’ appointment at Uppsala only following Ostwald’s vocal
interest in him for a Riga appointment. After 1884 none of the Ionists “could claim that
they were working in the intellectual vacuum that had surrounded each but a year
before.”54
The event that signified that the Ionists’ research could count itself part of a
community was the creation of the Zeitschrift für physikalische Chemie. In the previous
years the literature on physico-chemical topics had grown rapidly, “parceled out among
journals in almost every country and language in Europe as well as the United States and
England.”55 By now a professor in physical chemistry at the University of Leipzig,
Ostwald spearheaded the creation of the journal based on the support he had garnered
earlier in the decade. Growing well into his role as a skilled “clarifier and systematizer of
knowledge, and an indefatigable defender of the newer physical chemistry”, he felt that
there was now a need for a centralized location to publish new works in solution chemistry
and thermodynamics.56 Van’t Hoff was invited to be an editor, German was set as the
language of print, and in 1887, three years after Arrhenius’s dissertation and Van’t Hoff’s
53 Root-Bernstein, The Ionists, v. 54 Root-Bernstein, The Ionists, 172. 55 Root-Bernstein, The Ionists, 352. 56 Partington, A History of Chemistry, 597.
23
groundbreaking Études, the first issue was published. Despite some language barriers
created by the German requirement, the Zeitschrift attracted an international audience, and
those drawn in to associate with it “differed significantly from the scientific communities
that were linked to other publications in the physical sciences at this time.”57
The political and professional connections the Ionists had made in the previous
years were now formalized with this journal. As given by Nye’s scheme, this forms a major
signifier of new disciplinary growth. Chemists and physicists interested in research on the
new physical chemistry were no longer scattered voices on the scientific periphery: with a
new place to publish, separate from existing journals in the physical sciences. In the first
issue, a foreword by Ostwald championed the issue as the beginning of a new chemistry, a
“chemistry of the future”, but as Root-Bernstein argues, it can also be viewed as not the
beginning of the first stage of disciplinary growth, but the culmination of it. The new
journal invited external recognition that a “context, even a recognized need, existed for the
theories [its authors] were about to announce”, and provided evidence of the political
changes that the Ionists had accomplished in the previous years.58
2. Nineteenth-century debates and the problem of the atom
The creation of physical chemistry as a discipline marked the beginning of a
formal movement where physico-chemical research could be officially practiced, but in
the nineteenth century, this growth occupied a largely separate path from the growth of
chemical bond research. The modern study of chemical bonding, as first developed in the
late 1920s following the discovery of wave mechanics, relies on a rigorous understanding
57 Root-Bernstein, The Ionists, 361. 58 Root-Bernstein, The Ionists, 365.
24
of the physical properties of the atom and its subatomic particles. To understand chemical
bonding is to understand the nature of the atom as a chemical entity. The atomic theory,
however, was certainly not new in the twentieth century, and was intensely debated as a
hypothesis through the mid- and late-nineteenth century, a period very closely to the
origin of physical chemistry and the rise of valence theory. In order to embrace the
chemical bond, chemists would need to embrace an essentially hypothetical entity. In this
section we see that in the nineteenth century, valency research was limited by a prevalent
attitude that the atom was not universally seen as an acceptable object for chemical study.
Nineteenth-century attitudes toward atomism were largely shaped by the work of
John Dalton, who, in his A New System of Chemical Philosophy (1808-1810) laid the
foundation for the modern chemical atomic theory. Dalton’s System was a treatise on the
composition of chemical substances, and introduced several important concepts into the
discipline. The law of multiple proportions, which stated that elements combine
chemically always in fixed ratios of weight, allowed atoms to be defined not as fixed
entities but in terms of their role in chemical combination. An atom was regarded as the
smallest quantity that would produce a compound, and a molecule as the smallest that
existed within a compound.59 In addition, Dalton developed a symbolic notation, in
which each element was represented by a circle with a different distinguishing character.
These symbols were assembled to create pictorial representations of molecules, as a
symbolic representation of the proportions in which the elements combined.60
59 Rocke, Chemical Atomism in the Nineteenth Century, 207; Levere, Transforming Matter, 108-110. 60 Ursula Klein places Dalton’s work in the context of nineteenth century chemistry as a “paper tool” used in complement with experimental work, in Experiments, Models, Paper Tools: Cultures of Organic Chemistry in the Nineteenth Century (Stanford University Press, 2003). Gillian Gass extends this with a study of Dalton’s notation in “Spheres of Influence: Illustration, Notation, and John Dalton’s Conceptual Toolbox, 1803-1835”, Annals of Science, 63 (2007) (forthcoming).
25
Dalton’s theory gave chemists an ontological definition of the atom. This implied
a metaphysical acceptance of atoms preceded the investigation of their properties, in
contrast to the experimental definition of the earlier Lavoisier era, in which a simple
chemical substance was defined by the limits of experimental investigation. The atomic
hypothesis was assumed to be true in the Daltonian paradigm, in order to investigate not
the atoms themselves but their chemical properties. As chemists assumed axiomatically
that ultimate particles of substances were alike, and combined and re-combined in
reactions in fixed proportions of mass, they gained a useful set of principles to guide their
research.
Atomism in the Daltonian context was, unlike physical atomism, not a theory of
matter but a theory of its properties, which offered a great practical improvement to
nineteenth-century chemists. Atomic weights had been proposed before Dalton’s time,
but “[n]o certainty could be attached” to them, “since for a long time no method was
agreed for determining the relation between them and equivalents.”61 His atoms
functioned as “mental representations”62 that guided chemists through a great ‘make-
work’ programme of research which had the specific goal of determining the proportions
in which elementary substances combine. At the same time, John Merz has argued, the
atom became problematic for chemists because it raised new questions they had not
previously encountered. Into the mid-century there came queries of how such atoms
would be arranged in space, and why did some elements combine in more than one
proportion, with various other elements? Dalton’s simplification had to “give way to a
series of modifications which have rendered the atomic theory one of the most
61 Russell, The History of Valency, 8. 62 Merz, A History of European Scientific Thought, 394.
26
complicated machineries ever introduced into science.”63 Still unequipped to analyze
chemical behaviour at the atomic level, chemists in the post-Dalton era could see these as
metaphysical questions only.
Modern historians of chemistry have portrayed the year 1860 as a crucial turning
point for atomism, this being the year of an international meeting of 140 chemists at
Karlsruhe, Germany, where the relevance of the atomic theory for chemistry became a
subject of debate. The meeting exemplifies W.H. Brock’s argument that the “view that
Dalton immediately and successfully convinced chemists of the truth and power of
atomism has proved far too simple”, as in 1860 chemists argued for a clear definition of
the difference between an atom and a molecule.64 Alan Rocke suggests in contrast to
Brock that chemical atomism remained alive, well, and uncontroversial through the
nineteenth-century. For many chemists, it was subscribing to physical atomism that was
most challenging, since it required a commitment to the reality of atoms as fundamental
particles. The chemists at this meeting knew their interests in the atom were not the same
as physicists’ and this awareness raised questions about the purpose of their work. Brock,
Nye, and Rocke’s studies show that the metaphysical assumptions Dalton’s theory
required were now becoming problematic, even as its practical value became clear.
Chemical scepticism about the physical atom and the future of the chemical one
“became more open and vocal” after Karlsruhe, with August Kekulé lending a leading
voice.65 He was concerned that the terms ‘atom’ and ‘molecule’ be made precise for
chemistry so that chemical molecules not be confused with physical ones. He cited the
63 Merz, A History of European Scientific Thought, 397. 64 W.H. Brock, The Atomic Debates: Brodie and the Rejection of the Atomic Theory (Leicester University Press, 1967), vii.; Mary Jo Nye (ed), The Question of the Atom: From the Karlsruhe Congress to the First Solvay Conference, 1800-1911 (Los Angeles: Tomash, 1984), xx. 65 Rocke, Chemical Atomism in the Nineteenth Century, xii.
27
observation that highly kinetic gas molecules experienced some dissociation in the
vapour phase, as an indicator of a greater physical complexity in the molecule. However,
he believed it would be disingenuous to force a corresponding chemical interpretation,
since no such complexity had been indicated through chemical means of investigation. It
was not up to chemists, he believed, to debate metaphysical questions of the reality of
atoms, but rather whether “the assumption of atoms [was] an hypothesis adapted to the
explanation of chemical phenomena.”66 It was important to Kekulé to create a chemical
system from chemical arguments, without giving physical data the same level of privilege
as chemical data. As Marcellin Berthelot believed, in agreement with his contemporaries,
while Daltonian equivalent weights were based on chemical analogies, atomic weights
were based on physical hypotheses, which were “speculative and metaphysical”67
Valency skirted many of the same challenges as atomism during this time, but
was not accompanied by the same level of vocal debate. The concept was generally
understood as the ability of a substance to combine with others, in many ways assisted by
Dalton’s symbolic notation which showed binary, ternary, quaternary and larger levels of
possible complexity. Compounds were known to combine in certain fixed ratios and
arrange themselves into common groupings, but these schemes rarely associated fixed
linkages between atoms as a cause of compositional patterns. Nye and Russell’s histories
have shown, however, that physical realism was the least crucial element of
representational schemes.68 Notation was seen as “a kind of shorthand”, which did not
always require the user to draw a comparison with any real object.69 The idea of a bond
66 Rocke, Chemical Atomism in the Nineteenth Century, 316, quoting Kekulé. 67 Nye, The Question of the Atom, 69. 68 Nye, From Chemical Philosophy to Theoretical Chemistry; Russell, A History of Valency. 69 Russell, A History of Valency, 42.
28
itself implied a physical connection between atoms (of which the metaphysical status was
already uncertain), that was beyond the scope of chemical investigation.
There were tentative connections to valency in some instances of notation. The
radical theory of Berzelius and Liebig in the 1820s and 1830s, was based on the
assumption that some chemical groupings, called radicals, worked as a single unit in
reactions and were believed to be unalterable. Their notation showed element or radical
symbols connected through dashes that were mean to represent the ability of each one to
join with another, but this was not meant to lead to any physical conclusions about the
structure of molecules or orientation of atoms.70 At first the representation was “little
more than an economy in words”; the attempt at realistic symbolism would come
gradually.71 For sceptics, representation was problematic because it was difficult to be
anti-realist about entities that had been pictured as billiard balls and piles of shot since
Dalton’s System. As with debates about the atomic hypothesis, a variety of perspectives
came to light as chemists recognized that multiple possibilities could be “superior to a
simple but wrong explanation.”72
For Kekulé, who was wary of letting physical concerns overtake chemical
reasoning, chemical formulae were best seen as “summaries of reactions”, in the way that
round brackets, which were introduced by the London Chemical Society in 1864, could
indicate chemical groups “that are not separated from each other during the course of a
chemical reaction.73 He drew a clear distinction between formulae that expressed
70 Rocke, Chemical Atomism in the Nineteenth Century, 251. 71 Russell, A History of Valency, 42. 72 Nye, From Chemical Philosophy to Theoretical Chemistry, 72. 73 Russell, A History of Valency, on 70 and 64.
29
structure, and those that were concerned with reactivity, and emphasised the limits of
chemical investigation when it came to determining physical structure:
…the manner in which the atoms of the substance undergo change and destruction cannot possibly prove how they are arranged in an unreacting compound that remains unattacked. At all events it must now indeed be held as a task of natural science to ascertain the constitution of matter, and therefore, if we can, the position of the atoms; but this cannot be attained by the study of chemical changes, but only through the comparative study of physical properties of unreacting compounds.74
The chemical relationship between elements in a reaction could be accurately reflected in
notation and the growing concept of valency, but these did not speak to the structure or
arrangement of atoms in a molecule.
It was the idea of atoms having a definite relationship to each other that had the
potential to lead chemists into physical territory, and made the prospect of valency
studies more challenging. Edward Frankland, who was responsible for coining the term
“bond” in 1866 in relation to chemical structure, was hesitant to give it any significance
beyond that of chemical valency:
By the term bond, I intend merely to give a more concrete expression to what has received various names from different chemists, such as atomicity, an atomic power… I do not intend to convey the idea of any material connection between the elements of a compound…75
Ostwald student and Leipzig graduate T.W. Richards was likewise unequivocal, calling
the notion of bonds “twaddle”: a “mode of representing not an explanation.”76
There was a mixture of distrust of the atomic theory and an appreciation of its
usefulness. Some skeptics were hesitant to embrace the atom for metaphysical reasons.
The least aggressive of these was a kind of tentative agnosticism, characterized by what 74 Russell, A History of Valency, quoting Kekule, 143-4. Emphasis added. 75 Russell, A History of Valency, quoting Frankland, 90. Emphasis added. 76 Nye, From Chemical Philosophy to Theoretical Chemistry, 121
30
Brock has called the ‘textbook tradition’. These chemists used the term ‘atom’ and
included Dalton’s theories in their teachings and textbooks, but still “refused to commit
themselves to the real existence of such bodies.”77 Many simply adopted atoms as
expository or heuristic devices without confronting their physical nature.78 Influential
French scholar August Comte believed the atomic theory was a “happy generalization”
that improved chemical thinking, whether or not atoms were real.79
Straightforward positivists denied their existence entirely, and although the
conceptual framework of Daltonism persisted in publications, “[v]ivid mental pictures of
Daltonian atoms often found no place in the thoughts of chemists whose literature was
filled with suggestions that atoms did not exist, or if they did were not important. From
the position that the atomic theory was only a hypothesis so long as it was not possible to
observe an atom, positivists rejected using the atom as a causal explanation for chemical
phenomena. With the climate of skepticism that the physical atom could not reflect the
needs of chemists, and no widespread commitment to a realistic belief in Daltonian
atoms, there was no universal agreement on the relevance of atoms for chemistry.
One of the strongest opponents of the atomic theory in chemistry was the British
chemist Benjamin C. Brodie, whose debates on the subject with Alexander Williamson at
the Chemical Society in London were closely observed by Kekulé and others. Brodie
distrusted the atomic hypothesis for the effect it might have on chemistry as a discipline,
which he thought had become too accustomed to accepting hypotheses without
questioning or testing them. Dalton’s work seemed to him the “actual theory” of
77 Brock, The Atomic Debates, 10. 78 Mary Jo Nye, “The Nineteenth-Century Atomic Debate and the Dilemma of an ‘Indifferent Hypothesis’”, Studies in the History and Philosophy of Science, 7 (1976), 245-268. 79 See Brock, The Atomic Debates, in particular the Appendix, “Comte, Williamson, and Brodie.”
31
chemistry80 because it connected quantitative chemical analysis with the principle of
fixed proportions, but the symbols Dalton had used to represent the material differences
between atoms had made the theory too dangerous because they implied assumptions
about the atomic hypothesis that not all chemists adhered to. Atomism had become so
connected with the law of multiple proportions that by the 1860s chemical symbols could
not “well be separated from the consideration of the hypothesis which is expressed in
them.”81
In an effort to ground his science on a more positivist foundation, Brodie
developed what he called a “symbolic calculus”, a system of notation to represent what
was known about chemical composition, along the lines of Dalton’s combining
proportions, without requiring the assumption of an atomic hypothesis. He believed this
type of system would “free the science of chemistry from the trammels imposed upon it
by accumulated hypotheses”, endowing it with an “exact and rational language.”82
Chemists could only trust their work as long as they trusted it was derived from positive
facts. What Brodie felt was unreliable, and philosophically dangerous, was to subscribe
to a symbolic system without considering the truth of it. To use a system as an analogy
could be unreliable if it was not questioned, but even more harmful was to use such a
system as a tool to understand what we see. Brodie’s fear was that chemical research
would become bogged down by conjectures. Chemists needed to focus on the facts they
knew, and not hypotheses.
80 Benjamin C. Brodie, “The Calculus of Chemical operations; being a Method for the Investigation, by means of Symbols, of the Laws of the Distribution of Weight in Chemical Change. – Part I. On the Construction of Chemical Symbols,” Philsophical Transactions of the Royal Society of London, Vol 156 (1866): 781-859, on 782. 81 Brodie, “The Calculus of Chemical Operations”, 782. 82 Brodie, “The Calculus of Chemical Operations”, 788.
32
Ostwald shared this kind of scepticism. Like many continental scientists, he was
averse to subscribing to what seemed to be a mechanical doctrine, and atoms had become
lumped together with the undesirable aspects of a mechanical philosophy.83 Accepting
them would mean following in the tradition of scientific materialism, the view that matter
and motion were “the final principles to which natural phenomena in all their variety
must be referred.”84 Ostwald “could not rid himself of the feeling that atoms were only
“mental artifices”, the remnants of an irresponsible, speculative ‘Naturphilosophie”.85
Like Brodie, Ostwald believed it was dangerous to base research on hypotheses. Well
into the turn of the twentieth century, he maintained the belief that scientists should “give
up all hope of getting a clear idea of the physical world by referring phenomena to an
atomistic mechanics.”86
By the 1880s, most anti-atomists felt the atomic theory had lost any heuristic
value it could have had, and could no longer treat it as what Nye has called an
“indifferent” hypothesis.87 Ostwald’s response, shared with Ernst Mach and others in the
physical community, was to base his reasoning upon the theory of energetics instead of
atoms. Energetics was a programme where energy, not mass was the primary data.
Although atoms were not observable, energy and thermodynamic properties were
measurable and, to strong positivists like Ostwald, offered a sounder understanding of the
physico-chemical world. While they could not be sure of the nature of matter, they could
reinterpret material concepts as manifestations of energy; mass became the capacity
83 Nye, The Question of the Atom, xxiv. 84 Ostwald, Wilhelm “Emancipation from Scientific Materialism”, Science Progress, 4 (Feb. 1896) reprinted in Mary Jo Nye (ed), The Question of the Atom: From the Karlsruhe Congress to the First Solvay Conference, 1800-1911 (Los Angeles: Tomash, 1984), pp 377-354. 85 Hiebert and Körper, “Friedrich Wilhelm Ostwald”, 262. 86 Ostwald, “Emancipation from Scientific Materialism”, 346. 87 Nye, “The Nineteenth-Century Atomic Debates”, 256.
33
factor of kinetic energy, and volume and space “distance energy” appearing like
gravitation in a Newtonian framework. Ostwald hoped this programme would ensure a
“science free from hypothesis”, and free from reduction to mental fictions.88 These anti-
atomists strove to judge their science by measurable properties, since “all that is
necessary to know may be expressed in this manner.”89
In the long term, Ostwald’s anti-atomism had a limited tenure, but his views
influenced several of the physical chemists who followed him. At the height of his
enthusiasm, by 1904 Ostwald had been able to show that all stoichiometric laws could be
determined without giving recourse to the atomic hypothesis, by relying on Gibbs’
thermodynamic principles, van’t Hoff’s law of equilibrium, and Le Chatelier’s principle.
In a striking example of his students’ loyalty, T.W. Richards at Harvard long remained
committed to anti-atomism despite earning a Nobel Prize for research on atomic weights,
and University of Toronto chemist Lash Miller remained a sceptic well into the twentieth
century. Van’t Hoff, Ostwald’s disciplinary compatriot, despite promoting the tetrahedral
spatial arrangement of the carbon atom, admitted to believing that atoms, molecules, and
their shapes and sizes were still “something doubtful”.90 Arrhenius, by contrast, had
always believed in the physical existence of atoms, but fought an uphill battle to have
atomic research accepted as legitimate physico-chemical work, into the first decade of the
twentieth century.91
Following Jean Perrin’s experimental observations on Brownian motion in 1906,
which were regarded as empirical confirmation of the atomic theory, Ostwald finally
88 Ostwald, “Emancipiation from Scientific Materialsim”, 352. See also Nye, “The Nineteenth-Century Atomic Debates”, 264. 89 Ostwald, “Emancipation from Scientific Materialism”, 352. 90 Nye, “The Nineteenth-Century Atomic Debates”, footnote on 259, quoting van’t Hoff. 91 Crawford, “Arrhenius, the Atomic Hypothesis, and the 1908 Nobel Prizes”.
34
accepted a “granular conception of the structure of matter.”92 Arrhenius campaigned
actively for research by Ernest Rutherford, Max Planck, and other atomic researchers to
be awarded the Nobel Prize during the surrounding years, hoping to use the Nobel as a
political means of settling the debate, and it was on the strength of his convictions that
Ostwald at last conceded.93 While Mach never adopted an atomistic ideology, Ostwald
agreed, in his preface to the 1909 edition of his Grundriss der allegemeine Chemie, that
“the evidence for atoms was now overwhelming.”94 By the early twentieth century,
Ostwald’s positions on atomism and energetics were widely viewed as obsolete.
The difficulties facing atomism from chemical and physical sources make it
challenging to analyze the late nineteenth century as the conceptually formative period
for the chemical bond. The many controversial responses first suggest scepticism about
the introduction of physically-based hypotheses into chemistry, and the possibility of
separate chemical and physical viewpoints on the atom. Daltonian atomism had made this
possible, by dressing it in the valuable chemical framework of combining proportions and
relative weights, but the metaphysical assumptions that Dalton introduced eventually
became too problematic for the chemical atomic hypothesis to be exempt from doubt. As
Alan Rocke warns, the distinction between physical and chemical atomism must be
preserved to prevent making the naïve assumption that “the chemical theory was as
successful and as uncontroversial as the physical one was rejected.”95
Later, when the scepticism shifted from disciplinary concerns to metaphysical
ones, those who opposed the atomic theory could not be satisfied with agnosticism or
92 Hiebert and Körper, 464. 93 Crawford, “Arrhenius, the Atomic Hypothesis, and the 1908 Nobel Prizes”, 519. 94 Laidler, The World of Physical Chemistry, 142. 95 Rocke, Chemical Atomism in the Nineteenth Century, xii.
35
heuristic use of an hypothesis that could not be tested. By the late-nineteenth century,
when physical chemistry was finding its feet, “[t]here was no room for utilitarian
compromise”, or viewing the atom as simply a “logical useful artifice”; either one
accepted the existence of atoms despite the absence of proof, or rejected them on the
same grounds.96
Ostwald’s response is particularly challenging in light of his role in developing
physical chemistry. Through his positivist philosophy he served to strengthen the basis of
chemical theory. Through the discipline he founded he promoted the ideology that
chemical knowledge can be discovered through physical means, and formed within
chemistry a conceptual framework for cross-disciplinary work that would not have been
possible in early years of discipline-motivated atomic scepticism. However, as a staunch
anti-atomist, his beliefs coloured the physical chemistry community in such a way that
one cannot imagine the study of chemical bonding flourishing under his care.
The late nineteenth century was a complex period for the concepts of atomism
and valency, and raised multiple perspectives on how these theories should be related to
chemical research, and how much chemists should steer clear of lest they tread too far
into physical territory. Limited by first disciplinary and then metaphysical scepticism
about what chemists could know about reactions at the atomic level, these pursuits were
left to being “essentially intellectual”.97 In what Nye has characterized as a transition
from chemical philosophy to theoretical chemistry, tentative understandings of chemical
bonding were snagged on the ongoing atomic debates and resistance to incorporating
96 Brock, The Atomic Debates, 21. 97 Russell, A History of Valency, 313.
36
physical methods. Until chemists obtained theoretical tools to study them, valency and
atomism were limited to being philosophical models.
3. Growth of a new discipline: the institutional framework.
As we have seen, physico-chemical work in the late nineteenth-century was
closely tied to chemical thermodynamics, the study of chemical reactions at equilibrium
in the solution state, particularly in terms of the behaviour of ions in solution. This
research was not designed to determine structural information about chemical substances,
but rather to gain a general, foundational understanding of the reaction process from the
broad physical parameters of environments like the solution state.98 An early
understanding of valency from mid-century onwards was tempered by debate on the
reality of the chemical and physical atom, and a reluctance by many chemists to embrace
the atomic theory on disciplinary or philosophical grounds. Into the turn of the twentieth
century, then, as chemists lacked a way to study the precise properties of substances at
the atomic level, the chemical bond, like the atom, remained a purely metaphysical
concern.99
The area where the new science had the most profound disciplinary consequences
was in the institutional growth that accompanied the beginning of the Ionists’ movement.
Beginning with the creation of a school for physical chemistry at the University of
Leipzig, and spreading into other countries as Leipzig graduates extended the same
model into their home countries and institutions, physical chemistry enjoyed rapid
disciplinary growth around the turn of the twentieth century. Particularly in the United 98 John W. Servos, “G.N. Lewis, The Disciplinary Setting,” Journal of Chemical Education, 61 (1984), 5-10, on 6. 99 Russell, A History of Valency, 314.
37
States, the creation of new schools of research around the Ionists’ programme developed
a formal research and laboratory framework, spaces dedicated to the application of
physical methods to chemical problems. Although these laboratories were devote most
specifically to the Ionists’ research concerns, this disciplinary growth was an important
pre-cursor to chemical bond research, in allowing physico-chemical research to become
institutionally mature.
The first laboratory of physical chemistry was housed at the University of
Leipzig, under Ostwald’s direction, in the “physically unimpressive” facility at a former
carriage house.100 Ostwald had many challenges to overcome in the first few years of
operation: in his first year he had three assistants (Julius Wagner, an analytical chemist,
Ernst Beckman, of pharmacy, and Walther Nernst, who within a few years was directing
his own physical chemistry program at the University of Göttingen), and only three
students enrolled in the laboratory course during the 1887-8 academic year.101 The new
laboratory also faced the imposing competition of a well-established organic chemistry
tradition. Ostwald, meanwhile, had to build a new tradition for his science from scratch.
The new journal had created a place for publication, and expanded the international
audience for physico-chemical research, but at a university where no physico-chemical
tradition existed, Ostwald had to “equip a laboratory, win the respect of his colleages, and
recruit assistants and students”.102 In the earliest years of the laboratory Ostwald’s small
100 Servos, Physical Chemistry from Ostwald to Pauling 101 For a comprehensive biography of Nernst and his role in early physical chemistry, see Diana Barkan, Walther Nernst and the transition to modern physical science (Cambridge: Cambridge University Press, 1999), in particular Chapters 2 and 3 for his role in Ostwald’s laboratory and early contributions to the discipline. 102 Servos, Physical Chemistry from Ostwald to Pauling, 48.
38
group primarily collected experimental support for the electrolytic dissocation theory,
well aware they were working apart from the mainstream circles of chemical research.
Servos and Nye show that the laboratory of physical chemistry marked a much
clearer break from older traditions in chemistry than the community-building activity the
Zeitschrift had been. Although they made use of physical ideas and maintained contacts
in the physical community, their research “was not directed to questions that physicists
were asking”, and neither were they doing traditional chemical analysis.103 The
ideological connection shared by Arrhenius, van’t Hoff and Ostwald was a significant
discontinuity from previous modes of thought in chemistry. In the terms of Nye’s
scheme, chemists who wished to include physical methods in their research lacked the
elements that could define them within a specific tradition: the shared practices and
values, a common landscape, and scientific homeland that would allow them to be
identified as practising physical chemists. In the 1880s and 1890s the Ionists needed to be
set apart from the rest of chemistry and also from physics to ensure a viable career path
for themselves and their students.
Servos suggests the isolation of physical chemists was important in making their
work a distinct disciplinary practice. Ostwald was determined to make a contrast between
“the aims and methods of the new discipline” and those of organic chemists whom he
“characterized as powerful, hidebound, fact-mongering opponents of the young, ‘modern’
chemistry. With the research questions they were asking, however, the Ionists could not
help being the black sheep of chemistry. New physical chemists needed to be trained in a
new tradition, to be skilled in standard techniques but also have a command of calculus,
thermodynamics and kinetic theory. The laboratory at Leipzig created this opportunity. 103 Servos, Physical Chemistry from Ostwald to Pauling, 44.
39
By the 1890s Ostwald had started to attract a steady flock of students, and his program
amassed a strong international reputation. Many of his students had begun their academic
life as organic students, and had grown quickly discouraged by its competitive nature. By
contrast, physical chemistry was fresh, new, and afforded more professional opportunities
to chemists at the beginning of their careers.
One of Ostwald’s earliest students, and arguably one of his most successful, was
American chemist Arthur A. Noyes. Born in Newburyport, Massachussetts, Noyes was
interested in chemistry as a teenager, but could not afford to pursuer higher education
until a scholarship enabled him to enrol at the Massachussetts Institute of Technology
(M.I.T.) in 1883.104 After completing his Master’s in 1887 he was appointed Assistant in
Analytical Chemistry, but it was customary at this time for Americans to complete their
doctoral studies in Europe before obtaining a professorship. For Noyes, this had meant
going to Leipzig, but to study organic chemistry under the well-reputed Wislicenus. After
only a few months work, however, he became frustrated with organic studies,
discouraged by “the prospect of attaining significant results in a field in which so much
had been achieved, and so many were working.”105 Ostwald’s new program offered a
novel alternative to which students there were easily attracted, and Noyes was quick to
transfer to the physical chemistry group.106 His doctoral project was a study of deviations
from van’t Hoff’s law of equilibrium, as part of Ostwald’s general programme of
subjecting the new principles to quantitative testing.
104 Linus Pauling, “Arthur Amos Noyes: September 13, 1866 – June 3, 1936”, Biographical Memoirs of the National Academy of Sciences, 31(1947), 322-346. 105 Servos, Physical Chemistry from Ostwald to Pauling, 59. 106 Sherill, Miles S., “The Contributions of Arthur A. Noyes to Science”, Science, 84 (1936), 217-220, on 218.
40
Like many of Ostwald’s American pupils, Noyes was deeply influenced by his
time in Leipzig, both in terms of his research and the level of enthusiasm he showed for
the new science. “The record here is unambiguous”, Servos indicates, as Noyes wrote
home with notes of praise for his new supervisor.107 Students from M.I.T. and Harvard
formed a collectively strong presence at Leipzig, and eagerly embraced the liberal spirit
of combining physics with Chemistry.108 Americans came to German laboratories in the
1880s and 1890s because they offered better training in chemistry than in their home
laboratories, but Ostwald’s presented something more, a new disciplinary ideology and
an enthusiastic spirit. When students returned home they came back as the vanguard of
physical chemistry who would carry the new techniques to other schools, which could
now, in turn, be populated with physical chemists. When Noyes resumed his career at
M.I.T. he was poised to become part of a small, productive, elite group of American
chemists.109
As newly trained physical chemists like Noyes moved on from their doctoral
work, their challenge was to find positions where they could work as physical chemists,
which was at first difficult in the case of the German community. There, organic
chemistry still attracted much more financial support than physical chemistry, a gap that
widened during Ostwald’s tenure at Leipzig. Positions in the new field were more likely
to come as untenured lecturers or small professorships at the local Technische
Hochschulen, institutions that enjoyed “faster growth rates than the universities and more
open and egalitarian structures.”110 By the twentieth century, though, Leipzig was not
107 Servos, Physical Chemistry from Ostwald to Pauling, 63. 108 Servos, Physical Chemistry from Ostwald to Pauling, 61. 109 Daniel Kevles, The Physicists (fourth edition) (Harvard University Press, 1917), 154. 110 Servos, Physical Chemistry from Ostwald to Pauling, 51.
41
alone, as the Universities of Göttingen, Giessen, and Freiburg all had institutes of
physical chemistry, and by 1905 chairs had been set up at Berlin and Göttingen as well as
Leipzig. Institutions in other areas of the continent were not as receptive, according to
Servos. In France, German degrees and publications had less worth due to political
strains following the Franco-Prussian war, and the Ionists’ chemistry was a nearly
entirely German-language endeavour. In the United Kingdom, the first chairs of physical
chemistry, at Oxford and Cambridge, were not created until after World War I.
In contrast, physical chemistry grew much more rapidly within American
institutions. This was due to an academic climate more receptive to new growth which
had been made possible by developments in American science over the course of the late
nineteenth century.111 This scientific growth coincided very fortunately with the rise of
physical chemistry, and would eventually put the United States in a position to reward
Ostwald’s ambitions more richly than any other country.
American science in the post-Civil War era was a young, not yet mature
enterprise. While universities in Germany, France, and England had been well-
established as centres of learning, the comparatively younger United States was still
building its system of education into one that could support higher learning. Lacking elite
institutions, a foundation of professors and researchers, and a national curriculum that
embodied the importance of scientific research, Americans were far behind in forming a
community of academic elite, not just in terms of quality but also in size. In the 1870s
111 On the topic of the growth of American universities and American science during the late nineteenth century, please see Roger L. Geiger, To Advance Knowledge: The Growth of American Research Universities, 1900-1940 (New York: Oxford University Press, 1986), Daniel Kevles, “The Physics, Mathematics, and Chemistry Communities: A Comparative Analysis”, in Oleson and Ross, eds, The Organization of Knowledge in Modern America, 1860-1920, 139-172; Kevles, The Physicists, Robert E. Kohler, “The Ph.D. Machine: Building on the Collegiate Base”, Isis 81 (1990), 638-662; and John Servos, “Mathematics and the Physical Sciences in America, 1880-1930,” Isis 77:4 (1986), 611-629.
42
only about thirty chemists, twenty physicists, and even fewer mathematicians were
publishing research.112
The quality of research being done in America was also comparatively poor, and
motivated by democratic, Baconian fact-gathering with no abstract appreciation of ‘pure’
science: all data were given equal importance, “and so by extension were all data
gatherings.”113 Theoretical research was rare in the 1870s and 1880s, with few
practitioners to lead research programmes and insufficient mathematical training that
would encourage theoretical thinking skills to develop.114 Gibbs, who was one of the few
successful American theoreticians, worked largely in isolation and did not invite younger
physicists to participate in developing his work in the style of a research group.115 The
mathematical progress that was made was directed toward practical ends, such as refining
physical constants, being “more a servant than a queen” of the physical sciences.116 The
fact-gathering ideology meant that the pace of American science, through the 1880s “was
inversely proportional to its mathematical content.”117
The institutional framework to support academic science in America likewise
remained weak until about the 1890s. The opening of Johns Hopkins University in
Baltimore in 1876 has been seen as the beginning of a slow process of educational reform
in the country.118 Hopkins led the way to producing doctoral students in the 1870s and
1880s, began to encourage graduate research, and founded many academic journals, but
112 Kevles, “The Physics, Mathematics, and Chemistry Communities”, 140. 113 Kevles, The Physicists, 36; Kevles, “The Physics, Mathematics, and Chemistry Communities”, 147. 114 Servos, “Mathematics and the Physical Sciences in America.” 115 Kevles, The Physicists, 34. 116 Kevles, “The Physics, Mathematics, and Chemistry Communities”, 34. 117 Servos, “Mathematics and the Physical Sciences in America”, 612. 118 Kohler, “The Ph.D. Machine: Building on the Collegiate Base”, 639.
43
this same spirit was much rarer nationwide.119 State universities, which resulted from a
federal land-grant program at the beginning of the 1860s, tended toward more vocational
training with few specialized entrance requirements.120 But a liberal, elitist spirit grew
gradually along with more national educational reforms, and by the 1880s, to discuss
science meant being able to “mark oneself as a cultivated man.”121 The last few decades
of the nineteenth century were the first period where American scientists could begin to
count themselves as part of a broad community of researchers, even though the quality of
research lagged behind.
Especially for scientists, a great short-coming of American academia was the lack
of professional opportunities awaiting a successful Ph.D. This made it hard to entice
students into graduate research, and “[e]ffective Ph.D. programs existed at only a handful
of institutions.”122 It was not until the 1890s that the expansion of universities and the
need for professional academic positions began to work to the country’s benefit, creating
a supply-and-demand system where higher education could lead to a position teaching the
waves of students that were beginning to fill American schools.123 Around the time
Ostwald’s school was finally attracting a large student population, American schools
were still not able to offer a scientific Ph.D. of the same value, but there was at least an
academic framework to support the Americans who wanted a degree from their
homeland.
119 Geiger, To Advance Knowledge, 8. 120 Geiger, To Advance Knowledge, 6. 121 Kevles, The Physicists, 17. 122 Geiger, To Advance Knowledge, 27. See also Kevles, “The Physics, Mathematics, and Chemistry Communities”. 123 Kohler, “The Ph.D. Machine: Building on the Collegiate Base”, 641-2.
44
Noyes returned from Leipzig a changed man in 1890. M.I.T. was in need of
instructors and he was qualified and dedicated. After four years of teaching analytical and
organic chemistry he became assistant professor of physical chemistry in 1894, then
professor of theoretical chemistry in 1899. During these years he gradually began to
effect changes in the curriculum, so that by the mid-1890s, theoretical chemistry at M.I.T.
was essentially the Ionists’ physical chemistry. By the time he was finished with the
syllabus, more than half the hours of classwork were devoted to solution theory, the law
of mass action and electrolytic dissociation, with a third of those hours “devoted to the
study of chemical equlibria alone.”124 In these years Noyes also composed a series of
textbooks, most notably The General Principles of Physical Sciences (1902), and the
eventual culmination A Course of Study in Chemical Principles (1922), which were
designed to supplement laboratory instruction and repair some of the weaknesses in
theoretical instruction in the American physical sciences.125 In his texts Noyes promoted
a method of problem-solving where students would obtain a “concrete illustration” of
chemical laws and principles from laboratory results, then apply these principles in
reasoning through more abstract solutions.126 Through his teaching at M.I.T., Noyes
brought clarity to his subject, influencing a new generation of American physical
chemists.
Noyes’ most profound extension of the Ionists’ program in American was, in
addition to his teaching methods, the creation of a laboratory that emulated the system in
place at Leipzig.127 During the 1890s Noyes had been largely dependent on
124 Servos, Physical Chemistry from Ostwald to Pauling, 62. 125 Pauling, “Arthur Amos Noyes”; Sherill, “The Contributions of Arthur A. Noyes to Science”. 126 Sherill, “The Contributions of Arthur A. Noyes to Science”. 219. 127 Servos, Physical Chemistry from Ostwald to Pauling, 114.
45
undergraduates to carry out research for his studies, but by the 1900s enough American
students had been trained at Leipzig to warrant the creation of a similar laboratory in the
United States.128 In 1901 the M.I.T. board approved Noyes’ request to establish a
department devoted to research in physical chemistry – an endeavour that was half-
funded by Noyes himself – but erected only a temporary laboratory structure to support
it.129 It was only after Noyes was awarded a $2,000 grant from the Carnegie Foundation
that M.I.T. was fully convinced, but in December of 1903 the Research Laboratory of
Physical Chemistry opened with Noyes as director. He considered the venture to be at the
outset an “audacious proposition”, but felt that in being in charge of such a department
would realize “the highest ambition” of his life.130
The laboratory afforded all the benefits of Ostwald’s Leipzig institute, but in a
new American homeland, to serve as a new extension of the Ionists’ discipline. In
allowing for the pursuit of basic, ‘pure’ scientific research, M.I.T. advanced a step in the
pace of American science, and freed Noyes from many of his teaching burdens so he
could direct physical chemistry research full time. Financed through annual contributions
and run more as a personal institute than an institutional fixture, the M.I.T. laboratory
was “more like a German than an American institute”, and was designed to copy the
training process so many Americans were already experiencing abroad.131 Here, Noyes
set a high standard for American physical chemistry “that contributed to its rapid
progress to a pre-eminent position in the world.”132
128 Pauling, “Arthur Amos Noyes”, 325. 129 Pauling, “Arthur Amos Noyes”, 325; Servos, Physical Chemistry from Ostwald to Pauling, 111. 130 Sherill, “The Contributions of Arthur A. Noyes to Science”. 217. Quoting a letter from Noyes to a student dated November, 1901. 131 Geiger, To Advance Knowledge, 88. 132 Pauling, “Arthur Amos Noyes”, 326.
46
The creation of a new space dedicated to the Ionists’ brand of research ensured
more opportunities for American students in the new discipline, at the same level of
quality. Because of the success of this laboratory, and the success that physical chemists
enjoyed in the expanding academic climate of the 1890s and 1900s for many decades
“the bulk of the teaching of physical chemistry” in North American universities “was
done by professors who had either worked with Ostwald or had come less directly under
his influence.”133 After 1908, the vast majority of American physical chemists obtained
their PhDs from American schools: the work of chemists like Noyes made it possible for
American physical chemistry to stand apart from its German origins.134
By the early twentieth century, a series of developments in physico-chemical
sciences had created a disciplinary environment that made it possible for chemists to
make use of physical methods within the mainstream of their science. American chemists
in particular benefited from these changes, as they came at the apex of a period of
nationwide academic growth and reform. Concerns over atomic debates and the ability of
chemists to make claims regarding the real structure of molecules persisted through the
late nineteenth century, but by the twentieth, these became less problematic. As
American physical chemistry settled in the early decades of the twentieth century, a
physico-chemical research framework was in place to made chemical bond research more
and more of a reality.
133 Laidler, The World of Physical Chemistry, 29. 134 Servos, Physical Chemistry from Ostwald to Pauling
47
Chapter 2: Atomism, chemical autonomy and the valence bond:
Lewis, Langmuir and the cubic atom
“It is evident at last that physicist and chemist are observing the same atom and the chemist may be confident that his observations will be given more consideration in theories of atomic structure in the future than they have been given in the past.”135
1. Introduction
The early-twentieth century growth of the cubic model of the atom and the theory
of the valence bond, in the work of American physical chemists Gilbert N. Lewis and
Irving Langmuir, forms a significant historical break from previous traditions of valency
and the physical chemistry of the nineteenth century. Lewis’ cubic atom, which
postulated a static arrangement of valence electrons that allowed both the sharing and
transfer of electrons in bond formation, posed a notable contrast to the contemporary
physical model of the atom presented by Niels Bohr. My focus in this chapter will be to
examine the Lewis model and the factors that influenced its acceptance in the chemical
community. Here I argue that because of its worth as a model that could account for
bonding in both polar and non-polar substances, the cubic atom supported an autonomous
chemical perspective on atomism and chemical bonding during the late 1910s and early
1920s. Lewis and Langmuir presented a model of chemical bonding that attended to the
needs of chemists, and for this reason it was well-accepted for its explanatory, conceptual
and heuristic value well into the 1920s.
135 Worth H. Rodebush, “The Electron Theory of Valence,” Chemical Reviews, 5 (1928): 509-531, on 531.
48
Current history of Lewis, Langmuir, and the cubic atom is indebted to the work of
a small group of authors, primarily Robert Kohler and John Servos.136 A series of papers
by Kohler analyses Lewis’ professional relationship with Langmuir and the significance
of the cubic atom in the American and European chemical communities, and provides an
understanding of the contemporary response to Lewis’ work.137 Most recently, a special
issue of the Journal of Computational Chemistry was prepared on chemical bonding on
the ninetieth anniversary of Lewis’ inaugural valence bond paper.138 Articles by Sason
Shaik, Ana Simões, and Gernot Frenking and Andreas Krapp revisit Lewis’ research, his
process of discovery, and the value of his work in the context of modern chemistry.139
Simões suggests that Lewis’ presentation of his theory in the various fora of journals,
textbooks, and conferences allowed him to explore both sides of the physico-chemical
boundary, while Shaik argues for the valence bond’s value as part of a chemical
“heartland” territory, whose concepts remain important to chemical thought despite the
modern advances that have left Lewis’ original work obsolete. These more recent papers
present a general analysis of the original valence bond treatment for the modern chemist,
but offer only provisional historical insights into the development of his ideas, and lack a
comprehensive discussion of Langmuir’s contributions.
136 John W. Servos, “G.N. Lewis, The Disciplinary Setting,” Journal of Chemical Education, 61 (1984), 5-10, Servos, Physical Chemistry from Ostwald to Pauling. 137 Robert E. Kohler, “The Origin of G.N. Lewis’s Theory of the Shared Pair Bond,” Historical Studies in the Physical Sciences, 3 (1971), 343-376; Robert E. Kohler, “Irving Langmuir and the Octet Theory of Valence,” Historical Studies in the Physical Sciences, 4 (1974), 39-87; Robert E. Kohler, “The Lewis-Langmuir Theory of Valence and the Chemical Community, 1920-1928”, Historical Studies in the Physical Sciences, 6 (1975), 431-468. 138Gilbert N. Lewis, “The Atom and the Molecule,” Journal of the American Chemical Society 38 (1916), 762-785. 139 Sason Shaik, “The Lewis Legacy: The Chemical Bond – A Territory and Heartland of Chemistry”, Journal of Computational Chemistry 28(2007), 51-61; Ana Simões, “In Between Worlds: G.N. Lewis, the Shared Pair Bond and Its Multifarious Contexts”, Journal of Computational Chemistry, 28 (2007), 62-72; Gernot Frenking and Andreas Krapp, “Unicorns in the World of Chemical Bonding Models”, Journal of Computational Chemistry, 28 (2007), 15-24.
49
The current literature offers a tentative disciplinary analysis of Lewis as a
physical chemist receptive to new developments, and acknowledges the problematic
contrast between the Lewis and Bohr models of the atom. However, in the broader
context of disciplinary growth and the historical development of valence theory, the cubic
atom is much more valuable than as only a short-lived alternative to the physical one. As
a physical chemist Lewis was indeed well-versed in the physics of his time, and his
allegiance to a chemical model with physical flaws is a non-trivial component of his
guiding views. It is important that the chemical bond developed in the framework of a
chemical atomism, rather than as a wholesale appropriation of a physical mode of
thought. This perspective is absent from much of the current work, but is necessary to
place the chemical bond in its disciplinary context. In modern terms, Lewis’ model is
often seen as a historical relic that retains heuristic and pedagogical value, particularly in
modern chemistry classrooms. Analysing it in terms of Lewis’ goals and the needs
chemists and physicists had for an atomic model shows we must afford the cubic atom a
greater status as a signifier of interdisciplinary change in the pre-quantum mechanics era.
This chapter will show, through an examination of Lewis and Langmuir’s papers
and responses to the cubic atom that their allegiance to chemical goals limited the extent
to which the cubic model could fully account for the phenomenon of valence bonding. At
the same time, the development and acceptance of the model established an autonomous
chemical viewpoint on the atom and bonding. Conceptually and physically distinct from
the rival Bohr atom, the cubic model proved valuable to chemical reasoning because it
captured the key features of valence bond theory. It was these elements that helped the
cubic model become accepted, and what kept Lewis’ formulation distinct from
50
Langmuir’s extrapolation of it into the octet theory. Unlike the nineteenth-century
framework of atomic scepticism, it was possible in the early twentieth-century for
chemists to embrace an atomic model that spoke to their disciplinary needs.
The story of the early valence bond, then, is a story of an emerging physico-
chemical identity that differed from the experimental tradition created by the Ionists. Not
inspired by kinetic or electrolytic theory yet preceding the revolutionary changes of the
quantum-mechanical worldview, Lewis and Langmuir’s chemical bond occupies an
intermediary period in the development of the physico-chemical worldview. In this
period the valence bond optimistically offered a way to understand structural phenomena
for the first time in an atomistic mode. As we will also see in Chapter 3, with the
adoption of wave-mechanical methods, we see in the case of the cubic model that the
chemists who developed this theory looked to physics not to supplant chemical thought,
but to supplement it.
2. The cubic atom and early valence bond theory
2.1 Lewis’ academic influences
Lewis was born near Boston, Massachussetts, on October 25, 1875, where he
lived until the age of nine when his family moved to Lincoln, Nebraska.140 He had little
formal schooling as a youth, until enrolling in the university preparatory school at
Lincoln at age fourteen.141 After spending two years at the University of Nebraska, Lewis
transferred to Harvard and received a Bachelor of Science in Chemistry in 1896. In 1899 140 Biographical information on Lewis can be found in Joel Hildebrand, “Gilbert Newton Lewis: October 25, 1875-March 23, 1946”, Biographical Memoirs, National Academy of Sciences Vol 31 (1947), 210-235, Robert E. Kohler, “Lewis, Gilbert Newton”, Dictionary of Scientific Biography, Vol 8 (Charles Scribners & Sons: American Council of Learned Societies, Charles Coulson Gillispie, ed., 1973), 289-293., and Arthur Lachman, Borderland of the Unknown: The Life Story of Gilbert Newton Lewis (Pageant Press, 1955) 141 Hildebrand, “Gilbert Newton Lewis”, 210.
51
he completed his Ph.D. at Harvard under former Ostwald student Richards, and “made
the pilgrimage to Germany” the following year to study first with Nernst in Göttingen142
and then Ostwald at Leipzig. Upon his return he taught thermodynamics and
electrochemistry at Harvard from 1901 to 1904, then after a year’s leave joined Noyes’
group of physical chemists at M.I.T., where he was “rapidly promoted to a full
professorship.”143 Lewis remained happily at M.I.T. until 1912, when he took up the
Chair of the College of Chemistry at Berkeley, where he remained for the rest of his life.
His only extended interruption from academic life came during a research commission to
the Chemical Warfare Service in World War I from 1917 to 1918, for which he was
awarded the Distinguished Service Medal in 1922. A prolific researcher, Lewis published
on a range of topics such as electrochemical potential, acid-base reactions and
photochemistry. His year in the Chemical Warfare Service was the only year of his life in
which he had no publications.
The most ideologically formative period in Lewis’ career was his time at M.I.T.
and Harvard, where he received training in traditional physical chemistry from two
second-generation Ionists, Noyes and Richards. This began during his time teaching at
Harvard, when Richards was actively studying thermodynamics. He supervised a joint
paper with Lewis that incorporated the results of Lewis’ doctoral thesis, on the
electrochemical and thermochemical properties of zinc and cadmium amalgams. Richards
was an experimental perfectionist who worked closely with his graduate students, and
gave Lewis what Servos describes as “first and foremost a training in experimental
142 Kohler, “Lewis, Gilbert Newton”, Dictionary of Scientific Biography, Vol 8 (Charles Scribners & Sons: American Council of Learned Societies, Charles Coulson Gillispie, ed., 1973), 289-293., 289. 143 Lachman, Borderland of the Unknown, 23.
52
technique.”144 Both chemists learned from each other as they collaborated and taught
others, organizing and explaining the “scattered, unconnected, and often imprecise”
theoretical discoveries which then constituted the field.145
Despite this success, accounts of Lewis’ years at Harvard suggest the academic
climate there became a deterrent, which influenced Lewis’ decision not to return there in
1905. In Arthur Lachman’s recollection of the time he stated that while “Richards’
reputation remained outstanding”, the best that could be said of the remaining chemistry
faculty was that “they lived up to the Harvard ideal of that period of doing nothing, but
doing that like gentlemen.”146 In a letter Lewis wrote to Robert Millikan after being
settled at Berkeley, he expressed doubt about Harvard’s spirit of research. While he “had
very much the same ideas” on atomic structure then as he published after moving to
Berkeley, Lewis “could not find a soul sufficiently interested to hear it” at Harvard.147 As
seen in Chapter 1, Richards had also been heavily influenced by Ostwald’s anti-atomism
despite pursuing research into atomic weights. Branch states in his memoir of Lewis that
it was not clear why Lewis’ career at Harvard came to so soon an end, but it seems that
while these early years inspired a love of physical chemistry, it was not a supportive
environment for his theory of valency.148
The environment at M.I.T., by contrast, offered what was missing at Harvard.
Noyes’ group was “a band of talented physical chemists who exhibited not only the form
but also the spirit of research.”149 Richards had given Lewis superior laboratory training,
144 Servos, “G.N. Lewis, The Disciplinary Setting”, 7. 145 Servos, “G.N. Lewis, The Disciplinary Setting”, 7. 146 Lachman, Borderland of the Unknown, 21. 147 G.N. Lewis to Robert A. Millikan, 28 October 1919, quoted in Kohler, “The Origin of G.N. Lewis’ Theory.” 148 Branch, “Appendix: Gilbert Newton Lewis”, 18. 149 Servos, Physical Chemistry from Ostwald to Pauling, 119.
53
but Hildebrand suggests that M.I.T. was instead the place that most shaped the character
of his research. There, Lewis had seven years “marked by that intense scientific activity,
both experimental and theoretical, which continued throughout his entire career.”150
Unlike Harvard, M.I.T. offered a fresh curriculum moving toward the newest scientific
trend of physical chemistry, and was populated with young, enthusiastic researchers fresh
from German training. Noyes’ laboratory was a creative, active space where Lewis could
engage with the most progressive research in his field, including the problem of the
chemical bond.
When he moved to Berkeley, Lewis was offered much the same opportunity as
Noyes’ had, to build a new, modern, research group, and introduced changes in
curriculum and community-building that had lasting effects. When he began there the
College of chemistry “was badly in need of revitalization” and “Lewis was given
generous financial backing and a free hand in recruiting new faculty.”151 His reputation
had grown during his time at M.I.T. so that his presence at Berkeley was enough to
recommend the university to prospective graduate students, who were encouraged to go
to study with “the most brilliant young physical chemist at that time.”152 Lewis used his
administrative influence to maintain a high level of teaching, for undergraduate and
graduate courses, and to allow junior faculty to benefit from senior professors.153
Freshman chemistry courses soon grew to be “of enormous size”: at first more than 700
students, and grew to over 1500 after World War II.154 He opposed an over-specialized
150 Hildebrand, Joel, “Gilbert Newton Lewis”, 211. 151 Kohler, “The Origin of G.N. Lewis’s Theory,” 355. 152 Branch, “Appendix: Gilbert Newton Lewis,” 19. 153 Hildebrand reflects that “[t]he complaint that a freshman in a large university has no contact with professors has not applied in freshman chemistry at the University of California”. Hildebrand, “Gilbert Newton Lewis”, 213. 154 Lachman, Borderland of the Unknown, 44.
54
undergraduate curriculum in order to allow students to grasp the fundamentals of the
science first, and emphasized laboratory research for honours undergraduate and graduate
students. The practice of chemistry, under Lewis’ leadership, was the pursuit of pure
science, without immediate concern for its practical or industrial utility.
The department of chemistry at Berkeley became a success on a professional level
because of the quality of its faculty and graduate and undergraduate students. Just as with
the undergraduate curriculum, Lewis avoided over-categorizing his faculty into
subdivisions: all were titled ‘professor of chemistry’, and none were permitted to reserve
a field for their particular study. As a supervisor he was careful to limit the number of his
own students, to prevent “his eminence in chemistry from depriving the other members of
the staff of assistance in their work”, and “never destroyed his own usefulness” by over-
managing his laboratories.155 Berkeley soon became one of the most sought-after
destinations for visiting scholars and graduate students, achieving under Lewis’
administration a “magnetic drawing power.”156 Twenty percent of the National Research
Council Fellowships in Chemistry granted between 1909 and 1950 came to Lewis’ group.
Lachman recalls that though the number of doctorates granted in Lewis’ time was “not
really very large”, the significance of them was “in the quality of the graduates rather
than their number.”
The most celebrated component of Lewis’ leadership was in the Chemistry
department’s weekly research conference where Lewis, his graduate students and staff
were all educated about current research. As many as two hundred department members
would crowd into a smoke-filled room to hear reports on outside research, and then
155 Branch, “Appendix: Gilbert Newton Lewis,” 19; Lachman, Borderland of the Unknown, 42. 156 Lachman, Borderland of the Unknown, 48-9, and following. Emphasis added.
55
discussion of Berkeley laboratory work. Papers were followed by free, open inquiry “in
which everyone could have his say, irrespective of his rank or the orthodoxy of his
ideas.”157 Hildebrand and Lachman both reflect that this spirit carried over into the daily
interactions of the department members.
Any one who thought he had a bright idea rushed to try it out on a colleague. Groups of two or more could be seen every day in offices, before blackboards or even in the corridors, arguing vehemently about these “brain storms.” It is doubtful whether any paper ever emerged for publication that had not run the gauntlet of such criticism. The whole department thus because far greater than the sum of its individual members.158
It was a setting that challenged the participants to produce creative work of high quality.
Berkeley’s reputation grew because of the reputation Lewis brought to it, and the success
his leadership brought.
Like his former laboratory supervisor Noyes, Lewis cultivated and promoted an
active research and teaching environment in two universities that contributed to the
growth of American science in the early twentieth century, M.I.T. and U.C. Berkeley.
Though he did not create a school of research around chemical bond theory, Lewis’ work
was nonetheless influenced by the emerging community that transformed physical
chemistry from a specialized program into “the communal property of all chemists.”159
The disciplinary setting of theoretical chemistry, for Lewis, was not based around a
specific class of problems, as in the Ionist tradition, but in an open atmosphere that
promoted pure science and research growth through collaboration. His disciplinary
innovation came from situating the problems of physical chemistry as a part of general,
free chemical inquiry.
157 Branch, “Appendix: Gilbert Newton Lewis”, 20. 158 Hildebrand, “Gilbert Newton Lewis”, 492. 159 Servos, “G.N. Lewis, The Disciplinary Setting,” 5.
56
2.2 Lewis, the atom, and the molecule
Lewis’ earliest interest in the nature of the atom dates back to his time at Harvard
at the turn of the twentieth century. Still a student when Joseph John Thomson made his
first discoveries about the electron in 1897, Lewis was one of several new chemists who
grew curious about the electron’s chemical properties, despite a lingering climate of
atomic scepticism. In 1902, the year in which Noyes also published a textbook on the
principles of physical science with no mention of the electron,160 Lewis formulated a
memo in his notebook imagining what a chemical model of the electron in the atom
would look like; this rough sketch forms the first record of the cubic atom. As Kohler
notes, “few scraps of evidence have survived” to indicate what Lewis’ thoughts on
valency were during the years that followed, but he did initially find use for it as a
teaching device, to illustrate certain periodic properties of the elements in one of his
seminars.161
At the turn of the twentieth century, valence theory was still influenced by a
prevailing conflict of the nineteenth century. One line of thought, stemming from the
electrochemical studies of Swedish chemist Jöns Berzelius, held that chemical
combination was electrical in nature: elements were classified dualistically as either
electro-positive or electro-negative, and combined with an element of the opposite
type.162 This idea was generally suitable to treat inorganic substances and what would be
described in modern terms as ionic compounds, which exhibit stronger electrical
polarities and ability to ionize in solution. However, electrochemical dualism could not
well account for the properties of most organic compounds, which were generally
160 Servos, “G.N. Lewis, The Disciplinary Setting,” 9. 161 Kohler, “The Origin of G.N. Lewis’s Theory”, 352. 162 Russell, The History of Valency, 1971), 261-4.
57
electrically neutral, and appeared to combine elements of the same electro-positive or
electro-negative type in one molecule. With the rise of atomic scepticism in the later half
of the nineteenth century dualism declined, but chemists were divided along organic and
inorganic lines as to whether chemical combination was electrical in nature.
Thomson’s discovery of the electron supported the view that chemical valency
was electrical in nature, but the more important question for physicists was how to model
the structure of the atom based on the evidence of its positive- and negatively-charged
constituent particles. His original hypothesis at the turn of the twentieth century, known
as the plum-pudding model, at first placed electrons around the atom within a sphere of
positive charge, like the raisins in a pudding.163 Around the time Thomson first
developed this model, German electrochemist Richard Abegg discovered a law of
valence that showed the valences and counter-valences of each element added to eight,
which was advanced on the assumption of a polar framework for bonding.164 Bohr’s
model, first published in 1913, postulated electrons were placed in concentric rings
around a small, massive, positively-charged nucleus, following from Ernest Rutherford’s
discovery using radioactive emissions of alpha-rays that the bulk of an atom’s mass was
contained in the nucleus, and not in the negatively-charged area around it.165 In the
planetary model, as Bohr’s became known, electrons orbited the nucleus in the same way
163 J.J. Thomson, “On The Structure of the Atom: an Investigation of the Stability and Periods of Oscillation of a number of Corpuscles arranged at equal intervals around the Circumference of a Circle; with Application of the results to the Theory of Atomic Structure”, Philosophical Magazine, 7 (1904), 239-265. 164 Since elements exhibited more than one valence or power to combine, depending on what element they combined with, Abegg called these the principal and the ‘contra’ valences. These were also known as positive and negative valencies. For example, chlorine typically has a valence of 1 (e.g. when pairing with sodium to form NaCl), but can also have a contra-valence of 7 (e.g. in larger molecules like HClO3), and 1 + 7 = 8. W.H. Brock, The Norton History of Chemistry (New York: W.W. Norton & Co, 1993), 469. 165 Niels Bohr, “On the Constitution of Atoms and Molecules”, Philosophical Magazine, 26 (1913), 1 (part I), 476 (part II) and 857 (part III).
58
that planets orbited the sun, on fixed circular tracks spaced at distances according to
Planck’s quantum hypothesis.
By the time Bohr’s model was published, Thomson had also reconciled with the
quantization of energy. Thomson revised his model of the atom to place electrons in
concentric layers about the nucleus, according to how he could show that magnetic and
electromagnetic forces would keep them in balance.166 Quantization of energy was now
understood to be caused by the substructure of the atom, which became “the agent by
which the energy is transformed”.167 Following Bohr’s model was the ring atom of
German Walter Kossel, who suggested electrons were arranged in rings of eight,
postulating on the basis of X-ray absorption spectra that ionization preceded X-ray
emission.168 Kossel’s treatment, which combined the work of Bohr and English physicist
Henry Moseley, also influenced his Munich supervisor, Arnold Sommerfeld, while
Sommerfeld developed a modified Bohr model in which orbits became ellipses, and
quantum conditions for ionization were made more precise.
In these physical theories, learning about the process of chemical bonding was
secondary to simply modelling a stable atomic structure. Thomson’s work came the
closest of any to a treatment of valency on physical grounds, but his work still kept polar
and non-polar bonding in separate categories, or as he called it “two great classes” of
“charged” and “uncharged” molecules.169 The polar type formed a bond through simple
transference of ‘corpuscles of electricity’, in a process Thomson called intra-molecular
166 J.J. Thomson, “On the Structure of the Atom”, Philosophical Magazine, 26 (1913), 792-799; J.J. Thomson, The Electron in Chemistry (Philadelphia: The Franklin Institute, 1923). 167 Thomson, “On the Structure of the Atom” (1913), 792. 168 John L. Heilbron, “The Kossel-Sommerfeld Theory and the Ring Atom”, Isis, 58 (1967), 450-485. 169 J.J. Thomson, “The Forces Between Atoms and Chemical Affinity”, Philosophical Magazine, 27 (1914), 757-789.
59
ionization. He found this occurred only in cases where the bonding atoms were “well-
separated in the electro-chemical series” so as to create a strong difference in electrical
properties.170 Comparing both types of molecules, Thomson found they could be
differentiated by properties such as level of ionization in solution, and their condensation
in the gaseous state, but in general his best material came from the polar case. “Charged”
molecules simply exhibited properties that could be easily measured, because of their
pronounced electrical nature. However for the “uncharged” cases, Thomson, like many of
his contemporaries, had to hypothesize a physical cause for the chemical properties they
observed, without the benefit of any obvious reactive electrical signifiers. Physicists
needed more to work with, and so did chemists.
It was not until a decade after Lewis’ initial memo on the cubic atom, when other
members of the Berkeley group began to publish on the topic of valency, that the
polar/non-polar question was approached in broad chemical terms. William Bray had
been Lewis’ colleague at M.I.T. and had followed him to California when he began to
build the Chemistry department at Berkeley, while Gerald Branch was an English
chemist who joined Lewis’ new department.171 These two men vetted their ideas on
valency at the Berkeley research colloquia and published a brief paper in 1913 that raised
issues of how chemists should treat polar and non-polar substances in an electronic
framework. 172
170 Thomson, “The Forces Between Atom and Chemical Affinity” 171 Servos, Physical Chemistry from Ostwald to Pauling, 84. 172 William C. Bray, and Gerald E. K. Branch, “Valence and Tautomerism”, Journal of the American Chemical Society, 35 (1913), 1440-1447, on 1440.
60
In “Valence and Tautomerism”, Bray and Branch made a short study of the term
they called ‘valence number’, and applied it to the analysis of tautomeric compounds.
They believed that since “the evidence in favour of the existence of the electron and of
atoms and molecules” was then “as conclusive” as they could desire, chemists could now,
in 1913, admit to the “electronic conception of union” in both polar and non-polar
compounds.173 Bray and Branch suggested that the term valence “should be used only as
a number”, to indicate the number of possible sites of linkage for an atom in bonding, but
also believed that it expressed “the idea of the nature of the union between atoms.”174
Misunderstandings about the nature of polar substances, however, had led
chemists to confuse the concepts of polarity and valence number. Invoking a distinction
between polar and non-polar molecules along the historical lines of the different
structural and inorganic traditions, Bray and Branch proposed that both polar and non-
polar classifications could be adopted with the use of the terms “valence number” and
“polar number.” Polar number could be used to quantify the number of units of charge an
atom had lost during bonding “in an algebraic sense”.175 For example, although nitrogen
can have a valence of -3 in the polar case of ammonia because it attracts three hydrogen
atoms electrostatically, it also has a valence of 5 in the non-polar case of ammonium
chloride where it is linked to four hydrogens and one chlorine.176 Bray and Branch
believed assigning each atom both a valence number and a polar number would clarify
conflicting valence information and preserve the distinction between the two previous
173 Bray, and Branch, “Valence and Tautomerism”, 1442. 174 Bray, and Branch, “Valence and Tautomerism”, 1441, footnote 1. 175 Gilbert N. Lewis, “Valence and Tautomerism,” Journal of the American Chemical Society, 35 (1913),
1448-1455, on 1448. 176 Bray and Branch, “Valence and Tautomerism”, 1440.
61
modes of thought. As well, like Abegg, they had discovered a quantifiable property of
valence, but had not proposed any underlying physical cause for it.
Lewis’ first publication on valency followed Bray and Branch’s paper, in the
same issue of the Journal of the Amerian Chemical Society, and bore the same title.177
His own “Valence and Tautomerism” was a similarly succinct eight-page treatment of
valence theory in which he also sought to clarify the contemporary understanding of
valency and polar structures. He followed his colleagues’ definition of polar number and
valence number, but drew a much stronger distinction between the polar and non-polar
cases than Bray and Branch did. Cautioning that the distinction between polar and
valence numbers should not be lost, Lewis was adamant in 1913 that the existence must
be recognized of “two types of chemical combination which differ, not merely in degree
but in kind.”178 At this time, Lewis believed that polar and non-polar types were so
different as to be unrelated, extreme phenomena.
Lewis’ sharp distinction between the two types was initially motivated by the
different chemical properties they exhibited, which he believed were related to the nature
of the compounds’ electronic structure. Polar substances were more reactive, and more
readily ionisable in solution than their comparatively weak non-polar counterparts, so
Lewis hypothesised this was due to an underlying difference in electronic structure: the
absence of a “frame structure” in the molecule which would require the atoms to maintain
a single structural formula.179 Assuming this frame structure was present in the non-polar
compounds allowed Lewis to preserve the ideal of the structural tradition of valency,
177 See also Shaik, “The Lewis Legacy: The Chemical Bond”, 53-4 for a brief overview of Lewis’ first paper on valency. 178 Lewis, “Valence and Tautomerism”, 1448. Emphasis added. 179 Lewis, “Valence and Tautomerism”, 1451.
62
since “definit” (sic) points could still be assumed to be present, at which direct
connection to similar points on other atoms may be made. Bonding electrons of polar
bodies could move freely, which implied a collapse of the framework. These electrons
could then fall in and out of place “like the bits of glass in a kaleidoscope.”180 Lewis
further suggested the existence of a third, metallic type of compound, in which electrons
were so mobile they could move outside the molecule. Lewis further postulated the
existence of states of tautomeric equilibrium, between the polar and non-polar, and
between the polar and metallic states.
It is not clear from Lewis’ early publications whether or not he believed polar and
non-polar molecules had different ontological status based on their electronic properties.
He did believe that all the “distinguishing properties of the two types of compound” were
“necessary consequences” of the mobility of the electrons, but the “mere statement of the
polar number” could give no clues about the molecular structure while the non-polar case
could.181 For this reason he resisted employing notation that used arrows to indicate polar
electron transfer (which Bray and Branch did use), because of the limited structure
inherent in polar molecules. It is clear, however, that at this time Lewis left the actual
atomic structure arrangement behind the molecular structures as an open question, only
stating that a clear structure could be assumed only in the non-polar case. The importance
of this work was in simply confirming that a connection between molecular structure and
electronic arrangement could be made through a study of valency. The details of the
connection were still to come.
180 Lewis, “Valence and Tautomerism”, 1449. 181 Gilbert N. Lewis, “The Atom and the Molecule,” Journal of the American Chemical Society, 38 (1916), 762-785, on 764.
63
Three years later, Lewis presented a more sophisticated treatment of valency
using, for the first time in publication, the cubic model of the atom in the now classic
paper “The Atom and the Molecule.”182 He initially defined the cubic atom through six
postulates, three of which concerned the electronic structure of the atom, while the fourth
and fifth allowed guidelines for bonding.183 According to the first postulate, each atom
contained an unaltered “kernel”, which contains an excess positive charge corresponding
to the group the element belonged to in the periodic table. The full atom was composed
of this kernel and an external “shell”, which holds a number of electrons equal to the
positive charge of the kernel, leaving the whole body neutral. The shell held up to eight
electrons. Third, electrons in the shell could contain up to eight electrons, arranged
symmetrically at the corners of a cube. Fourth and fifth, the shells were mutually
interpenetrable, and although electrons might pass from one position to another in the
outer shell, they were held in position by constraints “determined by the nature of the
atom and of such other atoms” that combined with it.184 Lastly, the electric forces
between these very close particles do not obey Coulomb’s Law at short distances. The
choice of the cubic shape was motivated by the group of eight in the shell.
The fourth and fifth postulates formed the major components of Lewis’ theory of
valence, in combination with the concept of electron-pairing or the “rule of two”. He had
noticed in his years of laboratory experience a striking prevalence of molecules with an
even number of valence electrons, and a corresponding rarity of “odd” molecules, which
182 See also Shaik, “The Lewis Legacy: The Chemical Bond”, and Simões, “In Between Worlds”, for a brief overview of Lewis’ main points. 183 This paragraph and following summarize from Lewis, “The Atom and the Molecule”, 767-9. 184 Lewis, “The Atom and the Molecule”, 768.
64
suggested a natural tendency of electrons to pair up.185 Treating the atom as a cube
provided a straightforward way to let electrons pair, by the meeting of two atoms at a
cube edge or face. As the fourth and fifth postulate indicated that “in interpenetrable
atomic shells an electron does not belong exclusively to one single atom”, the joining of
cubes at an edge created a bond through the joint filling of two atoms’ electron shells.186
In the non-polar case the bond was a truly shared pair, but in the polar case, with these
electrons being extremely mobile, two atoms would link up electrostatically at an edge
following the transfer of an electron from one shell to another.187 In either type of
molecule the bond was regarded as a pair of electrons “coupled together… lying between
two atomic centres, and held jointly in the shells of two atoms.”188
One of the most remarkable features of this paper was that Lewis had reversed his
previous belief that polar and non-polar compounds were different natural kinds.
However, this change was not fully explained in this or any later publications. Kohler has
also stressed the change, but indicates there is very little surviving historical evidence of
the influences on Lewis’s thoughts, particularly from 1905 to 1913, but that Lewis still
felt “embarrassed about his radical change of tune” on this distinction.189 The transition
in thinking did have the important effect of unifying the two previous modes of thought
about valency. This will be important for our disciplinary story. The formal introduction
of the valence bond/electron pair no doubt was a part of his change of heart.
185 Lewis, Gilbert N., Valence and the Structure of Atoms and Molecules (American Chemical Society Monograph Series: The Chemical Catalog Company, New York, 1923), 81. 186 Simões, “In Between Worlds”, 66. 187 Lewis, “Valence and Tautomerism”, 1448-9. 188 Lewis, Valence and the Structure of Atoms and Molecules, 81-2. 189 Kohler, “The Origin of G.N. Lewis’s Theory”, 375.
65
Conceptually, the connection between the two types seemed an extension of his
earlier belief that they could exist in tautomeric equilibrium. By 1916, with the atomic
model as a common ontological basis for all chemical structures, Lewis could admit the
possibility of a gradual transition between the extreme cases. By “scanning the whole
field of chemical phenomena” and interpreting the variations in chemical behaviour, he
believed that chemists were forced “to the conclusion that the distinction between the
most extreme polar and nonpolar types is only one of degree”, not of kind.190 Robert
Paradowski has credited the insight of allowing the electron to become the common
property of two atomic shells with resolving Lewis’ earlier difficulties.191 It is likely that
when it became evident to Lewis that the cubic model explained both electronic
mechanisms, the behaviour of the bonding atoms seemed less distinct than previously
thought.
The cubic atom was a marked contrast from the physical model that was then
being developed following the work of Rutherford, Bohr, and Sommerfeld. The
similarities to other models, certainly, were present – the rule of eight as noted by Abegg,
the grouping of electrons in shells as suggested by Thomson and Bohr, and the positive
kernel that Rutherford discovered – but Lewis’ model embodied distinctively chemical
features of the atom, that could help chemists to understand the process of bonding.
Rather than the “purely schematic” role Abegg gave the rule of eight as a rule of valence,
for Lewis the octet was a “necessary consequence” of the cubic structure of the atom, and
the need for atoms to attain a complete valence shell through bonding.192 The goal of the
190 Lewis, “The Atom and the Molecule”, 773. 191 Robert Paradowski, The Structural Chemistry of Linus Pauling (Ph.D. Thesis, University of Wisconsin, 1972), 396-7. 192 Kohler, “The Origin of G.N. Lewis’s Theory”, 350-1.
66
cubic atom was not to describe the material, physical properties of chemical substances,
but rather the properties of the atom that could explain chemical combination.
This emphasis on chemical value was further exemplified by the notation and
definitions Lewis introduced in 1916. These emphasised the most chemically significant
portion of the atom, the valence or “shell” electrons, separate from the “kernel” which did
not affect the process of bonding. Lewis set the “common symbol of the element”193 in
bold text to represent the kernel, and the number of valence electrons separately with the
symbol E. Potassium hydroxide, for example was represented as KOHE8. To represent
the bonding pairs Lewis introduced pairs of dots surrounding the element symbol, such as
H : H for the single bond in H2. This notation could also represent the polarity of
different bonds in simple cases by introducing spacing: H :Cl, for example, indicated the
greater attraction of the bonding pair toward the heavily polar chlorine atom.194 These
symbols, like the cube-shaped atom, emphasised the structural properties of electrons in
molecules, determined by the chemist’s understanding of valency.
For Lewis, emphasising the chemical value of an atomic model did not mean
rejecting physical evidence about the atom, but rather limiting his claims to the things he
could be certain of based on chemical evidence. His proposed model did incorporate
some known physical atomic properties, such as in the sixth postulate, where he stated
that Coulomb’s Law was invalid at the small-scale distances of atomic combination. This
point formed the only purely physical constraint on Lewis’ system. He did not intend to
present an alternative physical theory to explain it, but simply make clear the physical
193 Lewis, “The Atom and the Molecule”, 768. See pp 768-772 for exposition. 194 Lewis, “The Atom and the Molecule”, 777-9.
67
limitations of the chemical system. Later in the paper he emphasised the importance of
building a chemical model using chemical facts:
Indeed it seems hardly likely that much progress can be made in the solution of the difficult problems relating to chemical combination by assigning in advance definite laws of force between the positive and negative constituents of an atom, and then on the basis of these laws building up mechanical models of the atom. We must first of all, from a study of chemical phenomena, learn the structure and the arrangement of the atoms…195
It was not his goal to construct a physical model or replicate the work that had already
been done by contemporary physicists. The best way to a theory of valency was to arrive
inductively at a chemical model of the atom based on chemical experience, rather than
deduce a model from the laws of physics.
3. Acceptance of the cubic atom
3.1 Langmuir and the octet theory of valence
Lewis’ theory of valence was ground-breaking in the early twentieth century
because it unified polar and non-polar types of molecules under a common theory of
valency. This made it “completely out of tune with established belief”, but the novelty of
Lewis’s ideas were not enough to afford the theory more than slow recognition outside of
the Berkeley community.196 Though he was an unmistakable presence in his local
department, Lewis was not as outgoing in presenting the cubic model in conferences and
other publications, and so it reached a limited audience. Much of the major exposition of
the theory was undertaken by Lewis’ American contemporary, the physical chemist
Irving Langmuir, whose development of the cubic atom into the octet theory of valence
195 Lewis, “The Atom and the Molecule”, 773. 196 Kohler, “The Origin of G.N. Lewis’ Theory”, 344.
68
rewarded the work with a wider audience, and gained wider acceptance for the chemical
model of the atom.
Langmuir was born in Brooklyn, New York, on January 31, 1881, and remained
in New York for his schooling years, with the exception of a three-year move to Paris for
his father’s work with New York Life Insurance. Until the age of fourteen Langmuir
“hated school and did poorly at it” but he became interested in science through his
brother Arthur’s encouragement, and soon dedicated himself to physics and chemistry. 197
After high school he received a degree from Columbia University in metallurgical
engineering in 1903.198 Like Noyes and Lewis he conducted doctoral work in Germany,
but by the time Langmuir ventured there Leipzig had become a far different environment
from the energetic place Noyes knew, and Ostwald rarely lectured.199 Preferring Nernst’s
laboratory and the promise of closer supervision, Langmuir completed a dissertation at
Göttingen on the dissociation of gases by wire filaments in 1906. His ambition upon
returning was to “spend his life at the frontier of physical-chemical research”, but he had
difficulty finding a research position at an American university.200 In 1909, after
spending three years teaching chemistry at a private boys’ school in Hoboken, New
Jersey, Langmuir was “thoroughly fed up” with teaching a dismal curriculum to unruly
students, and looked for a private research position if none were to be found in
ociety
cience and Technology in the Corporate
197 Hugh Taylor, “Irving Langmuir, 1881-1957”, Biographical Memoirs of the Fellows of the Royal S4 (1958), 167-184, on 168. 198 Kohler, “Irving Langmuir and the ‘Octet’ Theory”. Kohler suggests Langmuir chose engineering because it required a substantial physics and chemistry background. 199 Leonard S. Reich, “Irving Langmuir and the Pursuit of SEnvironment”, Technology and Culture, 24 (1983), 199-221, on 203. 200 Kohler, “Irving Langmuir and the ‘Octet’ Theory”, 42.
69
academia.201 He soon found one with the General Electric company in nearby
Schenectady, New York, where he remained for the rest of a long career.
In the early twentieth century General Electric (GE) was making a name for itself
as a company that was putting “pure” scientific research to work in an industrial
setting.202 In the late 1890s, GE’s German competitor, Allgemeine Elektrizitäts-
Gesellschaft (AEG), had successfully adopted a design for electrical filaments from
Nernst that improved effiency of their products by fifty percent, and when the same
design was bought by their American competitor George Westinghouse, GE made heavy
changes in company strategy to introduce pure scientific research into their laboratory in
the early 1900s.203 Rather than employ scientists as business analysts or product testers,
GE created a research environment for them that more closely resembled an academic
laboratory, such as Noyes’ research laboratory at M.I.T.
The company was well-attuned to the latest trends in physical chemistry and hired
an Ostwald graduate, Willis R. Whitney, in 1900 to direct the research group, for the very
practical reason that “ionists seemed to be very good at improving the incandescent
lamp.”204 Whitney was trained twice in the Ionist mold, having also received his
undergraduate degree at M.I.T. under Noyes’ supervision, and wanted to recreate the
spirit of free inquiry he had witnessed there and at Leipzig in his own research group.
201 Albert Rosenfeld, “The Quintessence of Irving Langmuir”, The Collected Works of Irving Langmuir, Volume Twelve (Pergamon Press, Inc, 1962), 3-229, on 67. 202 For the early history of General Electric and the direction of its research laboratory by Willis R. Whitney, see Laurence Ashley Hawkins, Adventure Into the Unknown: The first fifty years of the General Electric Research Laboratory (Morrow Press, 1950), Reich, “Irving Langmuir and the Pursuit of Science and Technology in the Corporate Environment”, George Wise, “Ionists in Industry: Physical Chemistry at General Electric, 1900-1915”, Isis 74 (1973), 6-21, George Wise, “A New Role for Professional Scientists in Industry: Industrial Research at General Electric, 1900-1916”, Technology and Culture 21 (1980), 408-429, and George Wise, Willis R. Whitney, General Electric, and the Origins of Industrial Research (Columbia University Press, 1985). 203 Wise, “Ionists in Industry, 12-13. 204 Wise, “Ionists in Industry”, 13.
70
Under his direction the research laboratory at GE exhibited a spirit of research that
Kohler has compared to Lewis’ department of chemistry at Berkeley.205 Whitney
encouraged his physical chemists to work freely on problems that sparked their interest,
without immediate concern for the industrial payoff. Open discussion was encouraged
through seminars, review sessions, and a weekly colloquia series as presented by GE staff
or invited speakers.206 By 1910 GE had become “one of the liveliest, most forward-
looking and productive research centers of chemical and physical research in
meric
on in the international community, from men
like Ne
A a”.207
For the first four years of his time at General Electric, Langmuir made a careful
study of tungsten filaments in light bulbs, to better understand the basic principles of
lamp operation and the “natural phenomena of the artificial environment.”208 He quickly
made a name for himself with this research, moving far beyond the simple incandescent
bulb to a theoretical study of electrical discharge in gases and in vacuuo, by examining
factors like the influence of temperature and pressure on the bulb, and the build-up of
electrical charge in the bulb apparatus. Ensuing studies of the heat lost from hot filaments
led him to the discovery that hydrogen molecules dissociated to free hydrogen atoms near
the filament, and a series of papers from 1912 to 1913 on the properties of atomic
hydrogen brought Langmuir wide recogniti
rnst, Rutherford, Lewis, and Bohr.
After he discovered Lewis’ work, the cubic atom seemed to Langmuir to offer
guidance along a possible line of inquiry into the processes he had been studying. His
205 Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, 44. 206 Wise, Willis R. Whitney, General Electric, and the Origins of Industrial Research 207 Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, 44. 208 Reich, “Irving Langmuir and the Pursuit of Science”, 210.
71
studies of gases intrigued Langmuir about the structure of matter, and he gradually
became convinced that the chemical forces between atoms and molecules were the same
as those that attracted molecules in states like films and surfaces. To Langmuir, the cubic
atom was simply another one of many steps made in the previous years toward an
understanding of the electric properties of matter. In his view the atom was not a purely
physical or chemical discovery. Lewis’ work, though, was special because it remedied a
deficiency of consideration to the “vast store of knowledge of chemical properties and
relationships” that could form a clearer foundation for an atomic model than the
“relatively meager experimental data” provided by physicists.209 The Lewis model was
more valuable to him than Bohr or Kossel’s because it could explain chemical properties
of matt
l
system
er.
Between 1919 and 1921, primarily, Langmuir worked on redeveloping Lewis’
theory into a more widely applicable theory of valence, using the cubic atom as a basis.
Lewis had established the importance of the group of eight electrons, but had applied it to
only a small range of compounds largely confined to the first two rows of the periodic
table. Langmuir’s goal was transform the theory into a more deductive, axiomatic system
that would give chemists (and physicists who were observing the progress) a genera
of valence, that could be applied from first principles to any chemical structure.
The major part of Langmuir’s exposition of the octet theory, “The Arrangement
of Electrons in Atoms and Molecules”, occupied more than twice the pages of Lewis’
two previous papers combined.210 Langmuir was rigorous in every place his predecessor
had been succinct, and saw himself as extending Lewis’ discoveries by filling in places
209 Irving Langmuir, “The Arrangement of Electrons in Atoms and Molecules”, Journal of the American Chemical Society, 41 (1919), 868-934, on 868. 210 Langmuir, “The Arrangement of Electrons in Atoms and Molecules”.
72
where there was an opportunity to expand. In this first paper he outlined eleven
postulates on the nature of the atom, beginning with seven specifically related to the
structure of the atom and its internal electronic arrangement. These followed from either
the known properties of the noble gases, upon which Lewis had based the stability of the
octet, and the atomic numbers given by the Rydberg series N = 2( 1 + 22 + 22 + 32 + 32 +
42 + …). Most importantly, he set apart the valence electrons from the rest of the atom,
since it was the “number and arrangement” of these particles determined “the ease with
which [atoms] are able to revert to more stable forms” through the sharing, giving, or
receivin
d before
moving on to bonding itself.2 the brief equation,
held in common by the octets. With this result Langmuir abstracted Lewis’ cube and rule
g of valence electrons of the atoms.211
In Langmuir’s hands the octet theory itself was essentially a mathematical
formalism designed to represent the facts of Lewis’ theory, and properties of atomic
structure. Langmuir worked hard to make the theory consistent with both chemical and
physical facts about the atom. It was only after cataloguing the 88 elements known at the
time in terms of the cubic model, and repairing the analysis that had been missing from
Lewis’ papers, that Langmuir began the discussion of valence. This was an effort to
ensure the atomic structure responsible for electron pairing was well understoo
12 The octet theory amounted to
e = 8n – 2p or p = ½(8n – e),
where e is the total number of valence electrons available in the bonding atoms, n is the
number of octets formed after bonding, and p is the total number of pairs of electrons
211 Langmuir, “The Arrangement of Electrons in Atoms and Molecules”, 873. 212 See Langmuir, “The Arrangement of Electrons in Atoms and Molecules”, 886-916, for his postulation of the octet theory in context of the properties of valence, and examples of the theory’s application.
73
of two to a simple calculation, reducing the process of bond formation to the placement
of valence electrons over the molecule as a whole.213
This expression for p, the number of active bonding pairs, allowed Langmuir to
clarify the location of covalent links in a molecule. In the simple non-polar molecule
CO2, for example, n = 3, and e = 4 + (2 x 6) = 16. Therefore for carbon dioxide p = ½(8 x
3 – 16) = 4, indicating the two carbon-oxygen links must be double bonds. The
phosphorus molecule, P4, was a particularly interesting case for Langmuir: the octet
theory indicated six bonding pairs (p = 6) between four molecules, which gave the two
structural possibilities of a square-shaped ring, or the linear arrangement P=P=P=P. From
this example Langmuir made a further study of phosophorus oxides. Throughout the
paper, octet theory calculations were accompanied by sketches of Lewis-esque cubes,
joined in the arrangements deduced from the derivations. Some allowances had to be
made in the application for polar compounds, which still could not be considered to
exhibit directional links as non-polar compounds did. In the case of NaCl, sodium loses
one electron to chlorine, but this was not considered to contribute a shared pair: for any
polar compound the number p must necessarily be zero. Therefore in NaCl, n = 1/8(e +
2p) = 1/8(8 + 0) = 1, making chlorine the only “active” octet in bonding.
What separated Langmuir’s work from Lewis’s was his discovery of several new
relationships between electronic structure and valency, and his introduction of new
terminology to describe special cases of bonding and electronic arrangement. Most
valuable was the discovery of isosterism, the phenomenon of two or more compounds
with the same number of valence electrons having similar physical properties despite
213 This approach later influenced Robert S. Mulliken’s development of molecular orbital theory, which treated the molecule as a whole rather than studying the individual links between atoms. Contrasts between the molecular orbital and valence bond approaches will be treated in later chapters.
74
different nuclear charges.214 For example, nitrous oxide (N2O), and carbon dioxide both
have sixteen valence electrons between three atoms, and exhibit similar electrical
conductivity, refractive index and magnetic susceptibility.215 Other isosteres included the
nitrogen and carbon monoxide molecules, and the ClO4 and SO42- ions. These similarities
exemplified the importance of the valence electrons. Langmuir used isosteres to
categorize ionic forms, and predicted similar crystal forms based on electronic
isomorphism.216 Further terms Langmuir introduced were “comolecule”, for a group of
atoms joined by pairs of electrons shared between adjacent atoms, and “covalence” was
the number of pairs of electrons any atom shared with adjacent ones.
Kohler argues that, due to Langmuir’s efforts, by 1923 Lewis and Langmuir’s
electronic theory of valence was well accepted by “a sizeable avant-garde in Britain and
America”, who applied it to the study of reaction mechanisms and molecular structure.
While not every chemist who used it believed in the realistic nature of a cube-shaped
atom, by the early 1920s it had become “too respectable a doctrine either to invite open
criticism or to need self-conscious support”.217 While most of the chemists who
researched chemical bonding at this time tended to be Berkeley graduates or Lewis’
colleagues, Langmuir’s speeches and papers were the main avenue for chemists to learn
the theory, and they successfully did so. In Germany, however, the situation was quite
different, because of a much stronger allegiance there to the Kossel atom, which lacked a
214 Irving Langmuir, “Isomorphism, Isosterism and Covalence”, Journal of the American Chemical Society, 41 (1919), 1543-1559. 215 Langmuir, “Isomorphism, Isosterism and Covalence”, 1544. 216 Langmuir, “Isomorphism, Isosterism and Covalence”, 1549, and surrounding pages. 217 Kohler, “The Lewis-Langmuir Theory of Valence”, 433-4.
75
conception of the “special stability of electron pairs”, and so did not promote the same
special chemical significance to the atom.218
With the octet theory, Langmuir moved past a re-interpretation of Lewis’ model
into a new formalism that could be used predictively to determine properties of molecules
based solely on electronic arrangement. While Simões sees Lewis’ original work as
following a “deductive style of presentation”, it is clear when contrasting the two
chemists’ work that Langmuir is the one who intended to create a deductive system of
valence theory.219 While Lewis’s postulates may have been presented in a similar
fashion, they served more as induction from chemical experience rather than the tools
with which to build a formal system. Kohler has described the octet theory as a
“rediscovery” of Lewis’ original work.220 In many ways it re-dressed the concepts Lewis
had already stated, but Langmuir’s work also advanced valence theory by formalizing
what Lewis did not. In many ways it embodied the aims of Lewis’ 1916 paper, to unify
many chemical structures under one theory, but in its deductive, formal nature, it was a
strong departure from the original discovery.
3.2 Priority dispute: Lewis and the reception of the octet theory
The differences between Lewis and Langmuir’s views precipitated a lengthy
dispute between the two chemists on questions of interpretation. A key problem for
Lewis was Langmuir’s emphasis of the group of eight over the electron pair, which the
older chemist viewed as the cornerstone of valence theory. By abstracting the process of
bonding to the placement of valence electrons in the context of the values e and n, the
218 Kohler, “The Lewis-Langmuir Theory of Valence”, 447. 219 Simões, “In Between Worlds”, 66. 220 Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, 64.
76
octet theory distanced the process of bonding too far from the cubic atom Lewis had
initially used as a bonding mechanism. To Lewis the electron pair was a naturally
occurring chemical phenomenon, which was the basis of chemical change. Unfortunately,
Langmuir’s clean formalism assigned no fundamental importance to electron pairing
other than a reference to the stability of the noble gas helium.221 The octet theory had also
lost Lewis’ characterization of polar and non-polar types. While a transfer of electrons
still constituted a case where the octet theory could be applied, the knowledge of whether
a bond was polar or non-polar still had to come from general chemical experience, and
was not captured axiomatically by Langmuir’s postulates.
The name “octet” theory was a similar point of dissatisfaction for Lewis, a
situation which conflated concerns of priority with those of over-emphasising the group
of eight. Langmuir had received an enormously positive response to his presentation of
the octet theory, and though he gave first credit to Lewis in every instance, audiences
responded more to Langmuir’s extroverted nature. His first audience at the American
Chemical Society meeting of April 1919, just a few months in advance of his paper
reaching print, responded so enthusiastically that he was invited to repeat his paper the
following day and lined up further visiting lecture opportunities on the topic.222 His
exposition became so effective that chemists who heard him speak were “swept off their
feet”, and the theory started to become known as the “Lewis-Langmuir theory” or even
associated with Langmuir alone.223
221 Langmuir, “The Arrangement of Electrons in Atoms and Molecules”, 888, Postulate Nine. 222 Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, 55. The paper was so successful that while at the conference Langmuir was offered a position in the physics department at Stanford University. Comfortable in the General Electric research laboratory, he declined. 223 Gillespie, R.J., and Robinson, E.A., “Gilbert N. Lewis and the Chemical Bond: The Electron Pair and the Octet Rule from 1916 to the Present Day”, Journal of Computational Chemistry, 28 (2007), 87-97, on 90; Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, 60-63.
77
In the priority dispute that followed, Lewis believed a source of the confusion was
Langmuir’s use of terminology that did not exactly represent Lewis’ original work. He
spoke of a “group of eight”, while Langmuir referred to an octet. Langmuir took priority
concerns seriously and made several concessions to Lewis, but grew frustrated by what
he saw as implications he had only rephrased Lewis’ work. In a long, diplomacy-laden
letter to Lewis, Langmuir pressed hard for the title “octet theory” because it represented
the principles of the theory instead of the people who created it. He wondered would it
not be more appropriate for it to be named the “Thomson-Stark-Bohr-Parson-Kossel-
Lewis-Langmuir” theory, it being simply one step in the work of many.224 “I think your
expression is preferable”, Lewis wrote in July of 1919, “but I should be very sorry to see
the whole theory known as the octet theory… especially because the octet is no more
fundamental to the theory than the electron pair which constitutes the chemical bond.”225
His concern was that as chemists were ascribing priority to Langmuir because of the octet
theory, they would also misinterpret the proper meaning of the theory as a whole.
The debate between Lewis and Langmuir indicates that the meaning and
interpretation of the octet theory of valence were as important as who should receive the
credit for it. While Langmuir felt he was preventing one chemist from taking priority
over a whole body of research, Lewis felt Langmuir’s approach misinterpreted the
original conception of the cubic atom. Kohler argues that many of Langmuir’s essential
developments had been part of Lewis’ first theory but since Lewis had not been explicit
about details, Langmuir believed he had made new progress. The “sense of revelation”
224 Letter from Langmuir to Lewis, April 1920, reprinted in Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, on 79. 225 Letter from Lewis to Langmuir, July 1919, reprinted in Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, on 74. Emphasis added.
78
for him was so strong, in an independent context, that for him “it was in fact a
rediscovery.”226 It is clear, when contrasting Langmuir’s rigorous system with Lewis’,
that their different styles heightened the misunderstandings of each others’ contributions.
Langmuir’s octet rule brought a new sophistication to the theory that Lewis had not
accomplished, and was more didactically effective in illustrating the value of the cubic
atom.
4. Emergence of chemical autonomy
Aided in large part by Langmuir’s success in popularizing Lewis’ model, by the
1920s the cubic atom was a success because it could explain the processes of bonding
and electronic properties of polar and non-polar compounds. In this period immediately
preceding the discovery of quantum mechanics, Lewis and Langmuir’s work promoted an
atomic worldview that spoke uniquely to the needs of chemists. Chemists of the late
1910s and early 1920s were no longer concerned with atomic scepticism. As physicists
and chemists now both began to model the atom, both disciplines approached a common
atomistic ontology because of common interest in the atom, but Lewis and Langmuir’s
treatments modeled uniquely chemical properties of the atom because of their primary
concern with describing the process of bonding. It will be seen in this section that
although the chemical atom embodied many physical flaws, the cubic atom’s value as a
model that could treat chemical bonding was evidence of an autonomous chemical view.
During this period, physics and chemistry remained separated by models of the atom that
reflected diverging disciplinary needs.
226 Kohler, “Irving Langmuir and the ‘Octet’ Theory of Valence”, 64.
79
As a physical chemist “descended” from Ostwald, Langmuir was very much
aware of the place his work had in the context of his discipline’s anti-atomistic roots.
Unlike many of his disciplinary ancestors, he believed the positivist scepticism of
Ostwald’s initial program of thermodynamics had “set back” the progress of
chemistry.227 Langmuir felt that the positivist tendencies of the late nineteenth century
had prevented chemists from taking advantage of a model that could explain the observed
phenomena. To him it was the absolutes of energy, temperature and entropy that were
“human inventions”, instead of the atom.228 To determine what model should be used to
describe a scientific phenomenon, the choice for him was not between fact and
hypothesis but between the concepts or models that could give scientists a better or worse
description of natural phenomena. Langmuir found that comparing Ostwald’s attempts to
practice and teach chemistry without use of the atomic theory with a modern course
showed him “a good understanding of what is meant by be tet r or worse.”
By 1920, the chief rival to the cubic atom was Bohr’s model, which was by this
time evidently flawed.229 Despite its successful connection to hydrogen line spectra, the
planetary model lacked a mechanical explanation for the orbiting electrons, to account for
how electrons could be in constant motion without losing momentum. This problem
remained unsolved until the development of quantum mechanics, but in the meantime
physicists were faced with employing an atomic model with properties they could not
fully explain.
227 Irving Langmuir, “Modern Concepts in Physics and Their Relation to Chemistry,” Science, 70 (Oct. 25, 1929), 385-396, on 393. 228 Langmuir, “Modern Concepts in Physics and Their Relation to Chemistry”, 393, and following quotation. 229 Helge Kragh, Quantum Generations: A History of Physics in the 20th Century (Princeton University Press: 1999), 157-9; Christa Jungnickel and Russell McCormmach, Intellectual Mastery of Nature: Theoretical Physics from Ohm to Einstein (2 vol.) (University of Chicago Press, 1986), vol. 2, 350-354.
80
In Lewis’ view the primary difference between the Bohr model and the cube was
the use of a dynamic or static electron, which heightened the disciplinary concerns he had
raised in his 1916 paper, of building a chemical model from chemically-valuable
evidence. Whether or not to accept a physical or chemical model, to him, came down to
the question of “whether the electrons in the atom and the molecule are at rest”.230 The
positions that atoms in organic, non-polar molecules in particular exhibited were so
definite and fixed that he found it inconceivable for the electrons that took part in
forming their structure could have any kind of motion.
More importantly, because there was no mechanics in place to account for the
orbital electron, there was no way of explaining how such an electron could take part in
bond formation, and Lewis was not willing to abandon his static cubic model for a flawed
physical one. Unless he and his contemporaries were willing, “under the onslaught of
quantum theories, to throw overboard all the basic principles of physical science”, they
had to admit there were non-negligible gaps in their understanding of the atomic
theory.231 For the purposes of explaining chemical combination, the best that could be
said was that a bond consisted of electrons held jointly between two atoms and
“constrained to definite positions by forces which we do not at present understand, but
which do not obey the simple law of inverse squares.”232 Since the cubic model allowed
him to model chemical bonding, Lewis saw no reason to reject it for its physically-flawed
competitor.
Despite these concerns about the physical model, the common physico-chemical
interest in the atom on the part of both chemists and physicists had led, by the early
230 Gilbert N. Lewis, “The Static Atom”, Science, 46 (Sep 28, 1917), 297-302, on 297. 231 Lewis, “The Static Atom”, 299. 232 Lewis, “The Static Atom”, 302.
81
1920s, to realise that the two disciplines were approaching common ground, and
repairing some of the previous disciplinary divisions. Thomson perceived that “ignorance
of the structures of the atom and molecule” had made the division between the sciences
feel more real, since there was no physical understanding of the inner workings of the
atom that could offer an explanation for the periodic differences in chemical properties of
the elements.233 However, now that the electron had been incorporated into physical and
chemical thought, Thomson found there was a much better understanding of both the
internal structure of the atom and “how one atom unites with others to form molecules”,
leading him to the conclusion that “the barrier which separated physics from chemistry”
was removed. Chemistry was, to Thomson, the area which would produce “the most
striking developments of the newer physics”, because of the “vast mass of information
accumulated by chemists” offered an “unrivalled means” of testing any conclusions about
the atomic theory. It was therefore only natural for the electron to dominate the future of
chemical theory.
Not all physicists shared Thomson’s optimism about allying with the path of
chemistry, but some, like Robert Millikan, believed the chemical model of the atom was
simply an alternative model of the same entity, directed toward explaining chemical
phenomena. Millikan agreed they had no answer to the arguments against the dynamic
model but, much like the chemists’ allegiance to their cubic model, saw no reason to
dismiss it so long as it was successful in accounting for known phenomena and had
predictive value. “For the present”, he said of the physical model in 1924, “it is truth and
no other theory of atomic structure need hope to be considered until it has shown itself
233 Thomson, The Electron in Chemistry, preface, and following quotations this paragraph.
82
able to approach it in fertility. I know of no competitor which is yet in sight.”234 In the
midst of these developments, the chemical atom still retained the electron as the
mechanism of bonding. Whatever might come in the future, Millikan was able to remark
confidently that the chemical atom would still remain “just what it has always been since
its discovery, the ultimate unit of chemical combination.”235
By the early 1920s chemists and physicists were aware that both the static and
dynamic models were flawed, and while there was an acceptance of a common atomistic
ontology there was a difficulty in reconciling physical and chemical interests in the atom
into a common model. Any chemical model had to explain, first and foremost, the
process of bonding, but the only model that satisfied this condition did so at the expense
of physical realism. Physicists, lacking this need, understandably favoured the Bohr
model that was designed to explain discoveries in their field. But gradually, in the years
immediately preceding the discovery of quantum mechanics, chemists and physicists
found their feeling of confidence in their models gradually eroding as they realized they
could not fully interpret the “structure and behaviour of the atom” without “essential
change in the modes of thought” which had before seemed wholly adequate.236
In 1923, with the publication of his major work Valence and the Structure of
Atoms and Molecules, Lewis sought to reconcile the two views, in synthesising chemical
and physical knowledge of the atom alongside a redressing of his 1916 theory of valence.
His goal was to “weld” existing physical and chemical discoveries “into a single view of
atomic structure”, while at the same time “bring to the better acquaintance of chemists
234 Robert Millikan, “Atomism in Modern Physics”, Journal of the Chemical Society, 125 (1924), 1405-1417, on 1414. 235 Millikan, “Atomism in Modern Physics”, 1414-15. 236 Lewis, Valence and the Structure of Atoms and Molecules, 18.
83
some of the astounding accomplishments of modern physics”.237 Situating his work as
simply one stage in a developing revolution in physico-chemical thought, Lewis hoped to
offer his chemical audience the best fit for an atomic worldview, according to what
chemists and physicists had learned about the atom.
This welding produced evident changes from the 1916 version of Lewis’ theory,
most notable of which was the absence of the cubic structure, which had been all but
abandoned in the face of overwhelming physical allegiance to the quantum theory.238
Based on the current body of physical evidence, Lewis proposed chemists “adopt the
whole of Bohr’s theory in so far as it pertains to a single atom which possesses a single
electron”, there being “no facts of chemistry which are opposed to this theory”, and now
admitted electrons should possess some orbital motion, with the energy difference
between orbits given by a multiple of Planck’s constant, h.239 However, the key
importance of the group of eight and the group of two remained unchanged: in Lewis’
1923 model valence shells still held up to only eight electrons, and the property of
electron pairing had “even greater significance” for him than it had in 1916.240
What had also gone unchanged was the absence of a physical explanation to
explain this electron pairing, which had been the fundamental motivation for retaining the
chemical atom. This missing component troubled Lewis much more than gradually
incorporating more physical postulates into his model, since the “tendency to form pairs
and the tendency to form groups of eight” were “the essential features in the arrangement
237 Lewis, Valence and the Structure of Atoms and Molecules, 42; Lewis, Valence and the Structure of Atoms and Molecules, preface. 238 Kohler, “The Lewis-Langmuit Theory of Valence”, 445. 239 Lewis, Valence and the Structure of Atoms and Molecules, 56. See pages 56-8 for full discussion of postulates taken from the physical theory. 240 Lewis, Valence and the Structure of Atoms and Molecules, 57.
84
of valence electrons in compound molecules.241 As late as 1925 the need for a physical
account of electron pairing was still “one of the few outstanding differences between the
physicist’s and the chemist’s views of the atom”, and Lewis was frustrated that physicists
had overlooked it for so long.242
In keeping with the goals of making the valence bond theory as widely applicable
as possible, Lewis spent the bulk of his discussion of the theory explaining how it could
be applied to unusual cases, such as double and triple bonds and exceptions to the rule of
eight. These cases made the theory more profitable because they could provide more
insight into the structural properties of unusual compounds and left open avenues for
further research.243 Boron compounds were particularly valuable examples, which
exhibited no more than three electron pairs in their valence shell. Compounds that had
more or less than four pairs of electrons in their valence shell exhibited unusual reactivity
that seemed to confirm the natural stability of the octet. These exceptions were
particularly exciting for Lewis, because “instead of weakening the fundamental theory,
[they] strengthen[ed] it.”244 They further confirmed to him that any attempt to deduce a
single rule of valence, as Langmuir had done, led to oversights. It was better to treat
molecules case by case.
Lewis had not been naïve in his reliance on a chemical model. He was aware of
the modern physical developments in atomic theory, and knew as well as any physicist
the weaknesses of the quantum theory. The attention he gave to the physical postulates of
his model indicates that if there had been a physical explanation of the electron pair, he
241 Lewis, Valence and the Structure of Atoms and Molecules, 66. 242 Lewis, Gilbert N., “The Magnetochemical Theory,” Chemical Reviews, 1 (1925), 231-245, 231. 243 See Chapter 8 of Valence and the Structure of Atoms and Molecules for detailed account of single and triple bonds, and examples in boron compounds. 244 Lewis, Valence and the Structure of Atoms and Molecules, 97.
85
would have willingly embraced a physical atomic model on behalf of his discipline. In
the absence of such a model, Lewis promoted a chemically autonomous view of the atom,
because it could meet chemists’ disciplinary needs. In his final words in Valence
cautioned chemists to allow experimental facts to guide further studies of the atom and
other unsolved phenomena. By maintaining an openness of mind, chemists would “not be
retarded by the conventions and inadequate mental abstractions of the past” when the
solutions did come.245
In the years before quantum mechanics, Lewis and his contemporaries were aware
of the constraints on their research into the atom, but conserved an optimism about the
research that lay in their future: if there were pieces of the puzzle still missing, the task
lay before them to find them. Thomson, unperturbed by the flaws in quantum theory, felt
certain his science would discover a “law of physics not recognized in the older science
which is all-important in connection with the theory of the atom and must form the basis
of that theory.”246 Lewis drew an analogy between the scientific situation and “that old
American institution, the circus”, where
the end of the performance finds the majority of spectators satiated with thrills and ready to return to more quiet pursuits. But there are always some who not only remain in their seats but make further payment to witness the even more blood-curdling feats of the supplementary performance.247
Lewis’ own show was coming to an end, but the supplementary performance would have
to be taken up by those chemists who would be able to wrestle with the “blood-curdling
feats” of the quantum theory, all the while remaining ever-mindful of the facts of
chemistry.
245 Lewis, Valence and the Structure of Atoms and Molecules, 163. 246 Thomson, The Electron in Chemistry, 5. 247 Lewis, Valence and the Structure of Atoms and Molecules, 162.
86
The short rise and fall of the cubic atom highlights the relationship between
valency and atomism, and physics and chemistry during the early twentieth century, at a
time when neither discipline had settled on a complete model of the atom. Although
Lewis was aware of the potential for physical theory to describe the behaviour of the
electron, he did not wish to only have physical theory naively inform chemical reasoning.
There was potential to unify the physical and chemical worldviews during this time, but it
would remain limited without a physical representation of the electron pair that preserved
the cubic model’s chemical value. It was not until the later developments of quantum
mechanics that chemists found what they needed from physics: a causal explanation of
the electron pair. It will be seen in the following chapter how this discovery filled the
needs of Lewis’ program.
87
Chapter 3: Seeds of emergence: wave mechanics meets the valence bond
“Bohr’s theory was unable to give a satisfactory understanding of the sharing of electrons in molecules, but the new quantum mechanics is showing itself capable of this.”248
1. Introduction
In Lewis’ hands, the valence bond proved to be a chemical work with an
undeniable connection to the physics of the atom. His work, and the work of Irving
Langmuir, showed there was potential for a study of chemical bonding to incorporate a
physical understanding of the properties of the electron yet still be guided by chemical
reasoning. What allowed this potential to be realized was the discovery, following the
development of quantum mechanics, that the physical property of resonance provided the
fundamental cause of the valence bond’s stability. This part of the story follows German
physicists Walter Heitler and Fritz London’s treatment of the resonance effect in the
hydrogen molecule, and its adaptation by American chemist Linus Pauling into a
programme of structural chemistry. Here I argue that Pauling’s use of Heitler and
London’s methods in the Nature of the Chemical Bond programme preserved the
tradition of valency Lewis began. In so doing, Pauling’s work re-imagined the valence
bond as a hybrid work, a physico-chemical synthesis of two previous traditions.
This chapter follows several areas of Pauling’s early career, from his discovery of
the valence bond problem as an undergraduate, to his doctoral training as a ‘modern’
physical chemist, and his postdoctoral research in Germany and ensuing research
programme in the chemical bond. In Munich he trained as a theoretical physicist and
248 Robert S. Mulliken, “Electronic Structures of Polyatomic Molecules and Valence”, Physical Review, 40 (1932), 55-62, on 54.
88
learned the new wave mechanics, before beginning his own research in the United States.
Part academic biography, and partly a story of personal discovery, this period from the
mid-1920s to mid-1930s exposed Pauling to a variety of disciplinary influences which
shaped the path of his career. This mixture of physical and chemical influences was
reflected in the character of the science he practiced. At the same time, we will see that
the use of highly technical, theoretical physical methods did not prevent Pauling from
allying himself in the community with Lewis’ tradition of chemical valency. As his
reputation in the ‘modern’ physical chemistry grew, he was still accepted in the chemical
community because he identified with their interests and need for a theory of valency.
2. Wave mechanics meets the chemical bond
2.1 Wave mechanics and the Heitler-London treatment
By the mid-1920s, it had become clear to physicists that the flaws in the Bohr
model of the atom were becoming insurmountable. While the hypothesis that electrons
were arranged in orbits (elliptical, quantized energy levels around the nucleus) was well-
supported by the line spectra of many known elements (hydrogen in particular), there also
seemed to be a “failure or anomaly” counted for each of the theory’s successes.249 The
planetary model of the atom was stable and predictively useful only for one-electron
systems such as hydrogen. By 1924 there was apparently no secure correlation between
experimental results and theoretical predictions, and “the accumulation of experimental
anomalies” had created a crisis in quantum physics.250 Theoretical physicists continued to
249 Kragh, Quantum Generations, 157-8. 250 Kragh, Quantum Generations, 159.
89
struggle with interpreting the modern discoveries of energy quanta, blackbody radiation,
and the behaviour of the electron in the atom, in terms of classical physics.251
From 1924 to 1926, a series of discoveries by European physicists led to a new
understanding of the physical world at the atomic level. These results rectified many of
the flaws of the previous framework. A major part of the debate concerned whether the
electron (as had been questioned about light) could be considered a particle or a wave. In
1924, Louis de Broglie proposed in his doctoral dissertation that particles can act like
waves, to account for an incompatibility in the tendency for relativistic energy to
increaste and the quantum energy to decrease for a particle set in motion according to
Einstein’s special theory of relativity. He accounted for the difference by suggesting a
pilot wave accompanied the particle, and suggested that both wave and particle theories
were acceptable for elementary particles. Following de Broglie’s discovery, Erwin
Schrödinger developed a mechanical treatment, in series of papers from 1925-6, that
could be applied to elementary particles, called wave mechanics. Here, the electron was
represented as a wave function ψ, where ψ is a solution to the partial differential equation
Hψ = Eψ. This result gave a relationship between the mass and energy of a moving
particle that was analogous to similar results in classical mechanics.
Just previously to Schrödinger’s discovery, German Werner Heisenberg, had also
developed a quantum mechanics founded on the relationships between only observable
quantities.252 Heisenberg’s positivistic mechanics was logically consistent and embodied
a spirit analogous to the familiar classical mechanics, but was also very abstract, and
251 See Thomas S. Kuhn, Black-body Theory and the Quantum Discontinuity 1894-1912 (Oxford University Press, 1978), especially 220-228, and for Planck’s role. 252 See Kragh, Quantum Generations, 161-3, and Jungnickel, and McCormmach, Vol 2, 363-5, in particular for the relationships between Born, Pauli, and Schrödinger.
90
embedded more in symbolic notation than visualizable physical entities. Heisenberg’s
collaborator Max Born realized his notation could be written as a calculus using matrices,
leading Born and Pascal Jordan to develop the famous commutation relation between
momentum and position:
pq – qp = h/2πi 1,
where 1 is the unit matrix. Heisenberg’s treatment introduced what became known as the
uncertainty principle, meaning it was impossible to know both the exact position and
momentum of a particle at the same time. This also introduced a great amount of
philosophical concern into the European physical community, for even as physicists
discovered more and more about the new quantum world, the more it seemed to be an
acausal, indeterministic one.
Schrödinger’s results were found to be more easily visualizable because the wave
function could be interpreted in terms of the spatial properties of the electron in the atom,
and it was soon recognized to be valuable for solving atomic problems.253 For chemical
problems, “which usually do not permit simple dynamical formulation in terms of nuclei
and electrons, but instead require to be treated with the aid of atomic and molecular
models”, Heisenberg’s matrix mechanics was unsuitable.254 Wave mechanics, the branch
of quantum mechanics that stems from Schrödinger’s equation, played for atomic
particles the same role that “classical mechanics plays for material objects.”255 In order to
constitute a solution, ψ must be a specific type of function, called an eigenfunction,
meaning that each quantum state or position of the electron corresponds to a unique
253 Jungnickel, and McCormmach, Vol. 2, 364. 254 Linus Pauling, “The Sizes of Ions and the Structure of Ionic Crystals”, Journal of the American Chemical Society, 49 (1927), 765-791, on 765, 255 Murrell, Kettle and Tedder, 3. For a derivation of the wave equation for modern students of chemistry, see the same, Chapter 2, “Matter Waves”.
91
eigenfunction.256 The term “orbital”, meaning a space “as much like an orbit as quantum
mechanics permits”, was given the meaning of the space determined by the solution.257
Due to the principle of uncertainty, the wave function was later given a probabilistic
interpretation by Born, where ψ 2ψ was given as the probability of the electron being in a
given location.
A further development that helped define the properties of the atom in the new
quantum-mechanical context was the postulation of a fourth quantum number. Late in
1924, Wolfgang Pauli proposed its existence after studying the Zeeman effect, in which
spectral lines split in the presence of a magnetic field.258 At this time he developed the
exclusion rule, which stated that no two electrons could share the same four values of
their quantum numbers. Soon after, in 1925, Danish physicists George Uhlenbeck and
Samuel Goudsmit gave this fourth degree the mechanical interpretation as spin. The
following year, they used the spinning electron to calculate the hydrogen spectrum. In
spatial terms, the discovery of spin and the exclusion rule now meant that if two electrons
were to share space in the same atomic orbital, they would need to have opposite spin.
This property, as will be seen in later sections, had direct relevance for the valence bond
treatment of the chemical bond.
As the reception of the Lewis atom had made evident, the chemical worth of an
atomic model was measured by its ability to account for chemical bonding, and
accordingly it was not until Schrödinger’s wave mechanics was applied to the valence
256 The partial derivatives of the wave equation constitute a quantity known as the Hamiltonian, given by H, in Hψ = Eψ. Because of this specific mathematical relationship between H, ψ and E, the energy of the particle, the function ψ is an eigenfunction and E an eigenvalue of the system. 257 Laidler, The World of Physical Chemistry, 345. 258 Michela Massimi, Pauli’s Exclusion Principle: the origin and validation of a scientific principle (Cambridge University Press, 2005), 32. See especially 72-77 for the timing of Uhlenbeck and Goudsmit’s discovery following
92
bond that chemists became aware of the new physics. This was done by two young
German physicists, Walter Heitler and Fritz London, in 1927.259 Their work began
simply as a collaboration of two physicists fresh from their graduate studies, but because
of the research it later inspired, became familiar by the late twentieth century to “every
undergraduate in physics and chemistry departments”, and “accepted as one of the
foundations of modern chemistry.”260
For Heitler, physics “didn’t exist” “as a subject of its own” in the early twentieth
century at his Karlsruhe Hochschule, so his first exposure to the problem of the atom had
been in his chemistry class.261 After deciding to pursue physics at the university level, he
reached Berlin only to find himself woefully under-prepared for the mathematics that
theoretical physics required. Heitler also found himself intimidated by the community
there, and the ‘great’ professors (Albert Einstein, Planck, Max von Laue, Nernst and Fritz
Haber) all too remote from the students. It was the University of Munich, rather, that was
known as “the place to go if you wanted to get your Ph.D.”,262 and so Heitler moved
there to pursue a doctoral project in statistical mechanics under Karl Herzfeld, not with
Arnold Sommerfeld as a theoretical physicist might have expected in hindsight.
London’s academic career had begun much earlier than Heitler’s, but in
philosophy, with a doctoral dissertation in 1921 on deductive systems. Resigning a
teaching career at a Berlin Gymnasium to work with Born, London was eventually
persuaded by Born to “do an actual calculation” and go to work with Arnold
259 Walter Heitler and Fritz London, “Wechselwirkung neutraler Atome und homopolare Bindung nach der Quantenmechanik”, Zeitschrift fur Physik, 44 (1927), 455-472. 260 Nevil Mott, “Walter Heinrich Heitler: 2 January 1904 – 15 November 1981”, Biographical Memoirs of the Fellows of the Royal Society, 28 (1982), 140-151, on 141; Russell, The History of Valency, 295. 261 Walter Heitler, Transcript of Interview with John L. Heilbron, March 18, 1963, Niels Bohr Library, AIP. 2. Heitler also felt introducing the atom “from the side of chemistry” was “of course, historically correct.” 262 Heitler, Interview with Heilbron, March 18, 1963, 3.
93
Sommerfeld’s theoretical research group in Munich.263 After publishing some work in
spectroscopy in 1925, and becoming Paul Ewald’s assistant in Stuttgart, London became
interested in the new quantum mechanics. He applied for a fellowship in 1926 from the
International Education Board (which later became the Rockefeller Foundation) and
eventually came to work in Zürich in April of 1927, to study wave mechanics under
Schrödinger.264 It was there that he met Heitler, and the two began their unexpected and
fortuitous collaboration.
When London and Heitler met in Zürich they were enthusiastic about
Schrödinger’s wave mechanics papers, as was “everyone” in the community.265 The two
young men had come hoping to participate in the new physics, under the supervision of
the now-legendary Schrödinger. They selected the research problem of bonding in the
non-polar molecule which, as seen in Chapter 2, had long been regarded as a distinct
natural process from polar bonding. Non-polar compounds, being electrically inert, had
been “extraordinarily hard to understand”266 under the older framework of quantum
theory: Heitler himself recalled years later that he could not remember a time when he
“thought that non-polar binding was solvable in terms of the old physics.”267 However,
quantum mechanics had given theoretical physics new possibilities, and if the mechanism
of bonding could be described in terms of the properties of electrons, Heitler and London
believed their new wave mechanical properties could be put to good use.
263 Kostas Gavroglu, Fritz London: A Scientific Biography (Cambridge University Press, 1995). 264 Gavroglu, Fritz London, 42. Pauling applied for the same fellowship in the same year, intending to work with Sommerfeld and Bohr. 265 According to Heitler, the Munich physics group received Schrödinger’s papers better than those of matrix mechanics because his writing was clear, the system was visualizable, and they dealt with “Sommerfeld’s favourite subject”, differential equations. Interview with Heilbron, March 18, 1963, 8. 266 Walter Heitler and Fritz London, “Wechselwirkung neutraler Atome”, on 455. Translation is my own. 267 Heitler, Interview with Heilbron, March 18, 1963, 16.
94
To study the physical cause of the valence bond, Heitler and London approached
the simple case of the hydrogen molecule to see what would happen when two hydrogen
atoms were brought into close proximity. At that time there had been “a very great deal
of discussion of the Pauli principle and how it could be interpreted”.268 And so when
Heitler and London set up the problem as simply “a useful application of Schrödinger’s
equation”, they initially considered only the interaction of the atoms and charge densities
of the bonded particles.269 At first they did not see any special connection between the
electron pair and any quantum effects. When they solved the wave equation for the
system, however, they found there were greater forces at work than Coulombic and Van
der Waals interation. While the energy of the single shared-pair bond in H2 did have a
strong Coulombic component, a much larger portion of the energy arose from the wave-
mechanical nature of the system. It was only after discussing the problem for several
weeks that they discovered the unknown portion arose from Heisenberg’s resonance
phenomenon, which manifested as “exchange energy” held by the bonding electrons.
After a day’s work they knew they had “the formation of the hydrogen molecule in [their]
hands”.270
Heitler and London’s treatment, which they outlined in their 1927 paper, set up a
mechanical system involving two nuclei, a and b, and two electrons, according to the
following scheme.271 Let ψ represent the wave function for an electron at nucleus a, and
φ represent the wave function for an electron at nucleus b, and let the two electrons be
268 Heitler, Interview with Heilbron, March 18, 1963, 16. 269 Mott, “Walter Heinrich Heitler”, 143. 270 Heitler, Interview with Heilbron, March 18, 1963, 16. 271 Heitler and London “Wechselwirkung neutraler Atome”. This sketch summarizes points from this paper.
95
identified by subscripts 1 and 2. It is possible to represent such a system by one of two
configurations depending on how the pair of electrons is oriented:
ψI = ψ1 φ2, or ψII = ψ2 φ1.
These configurations corresponded to the possibility that the electrons may change
places. Because electrons are indistinguishable from one another, both configurations
needed to be admitted as potential states for the molecule. Therefore, Heitler and London
represented the true state of the molecule as a linear combination of the two states,
ψ = a ψI + b ψII
for real numbers a and b.
While the paper was well-received at the 1927 meeting of the Deutsche
Physikalische Gesellschaft in Freiburg, the element of Heitler and London’s work that
drew the most attention, and later, controversy, was the meaning of the exchange energy.
At the meeting, Heitler and London were asked questions, particularly by experimental
chemists and physicists, of what exactly was exchanged in the process. Heitler and
London themselves saw the meaning of resonance as a known problem in quantum
mechanics that had to be seen as “a fundamentally new phenomenon that has no proper
analogy in the older physics.”272 Pauling later interpreted the resonance effect as
“involving an interchange in position of the two electrons forming the bond”.273, 274
Heitler believed the best way to describe the treatment was that “to each chemical
formula there corresponds a wave function”, but “the wave function is not such that it 272 Heitler, Interview with Heilbron, March 18, 1963, 18. 273 Linus Pauling, “The Shared Electron Bond”, Proceedings of the National Academy of Sciences, 14
(1928), 359-362, on 359. 274 For a further discussion of the history of the resonance controversy, particularly in the context of the reception of the valence bond and molecular orbital theories of bonding, see Park, Buhm Soon “Chemical translators: Pauling, Wheland and their strategies for teaching the theory of resonance”, British Journal for the History of Science, 32 (1999), 21-46, and Simões, and Gavroglu, “Issues in the History of Theoretical and Quantum Chemistry”.
96
corresponds to one chemical formula alone, but it is in general a combination of
several.”275 The most important outcome of Heitler and London’s study was the
discovery of resonance, a puzzling yet apparently typical phenomenon in quantum
mechanics, but most importantly, it provided the energy that stabilized the electron-pair
in the valence bond.
2.2 The physico-chemical community responds.
Physicists were inspired and challenged by new questions following the discovery
of quantum mechanics. Heitler and London’s work was part of a fresh trend of physical
research to look for applications of the new theory, and indeed the previously unsolved
mechanism of the shared-pair bond, which Lewis had postulated on the basis of chemical
evidence, could now be solved using wave mechanics. For chemists, it now became clear
that a physical atomic model could speak to their disciplinary needs. For physicists,
however, the importance of Heitler and London’s work lay in the introduction of new
problems of study, but whether the physicists’ handling of these problems would be in
the interests of chemists was open for debate. Resonance, which formed the basis of the
treatment, also served to open the problem of chemical bonding to more disciplinary
challenges, as the chemical and physical communities took differing interests in the re-
discovered valence bond.
Heitler and London’s discovery effectively changed what was previously known
as a chemical phenomenon into a product of a quantum-mechanical world. Resonance,
and the property of spin, the fourth quantum number, gave effective indirect support to
their work. According to the exclusion principle, two electrons could not share the same 275 Heitler, Interview with Heilbron, March 18, 1963, 21.
97
space, such as in the instance of overlapping bonding orbitals, unless they had the same
spin. This formalized in physical terms the concept of the electron pair: “…two electrons,
and no more, can enter into a bond, and once such a bond is formed the electrons
concerned in it can form no more bonds.”276 In a quantum-mechanical framework, not
only did the electron pair seem to be a naturally occurring physical phenomenon, the
“exchange of two electrons” in the valence bond was a fundamental phenomenon as
well.277
Further research on the wave-mechanical treatment of bonding showed just how
much of a challenge physicists could expect from this new research. While Heitler and
London were successful in solving the case of the hydrogen molecule in principle, more
complex systems than hydrogen were much more difficult to solve in practice. Showing
how something could be done proved to be easier than actually doing it: because of these
difficulties, approximations became heavily used to make practical solutions to the wave
equation possible. In 1933, Hubert James and A.S. Coolidge developed a completely ab
initio method of solving the wave function for diatomic hydrogen, that used a trial wave
function to predict the energy of the hydrogen bond with 98% accuracy. However, it took
over a year to complete the results for each term in the expansion of the wave function,
and they could not extend the method even to other light diatomic molecules.278
Currently, the most successful methods use an approach called configuration interaction
(CI), which relies on a basis set of single-electron wave functions taken in linear
combination to build up the molecule, in a series of stages, and produces a trial wave
function in an attempt to find a minimal energy solution to the Schrödinger equation. For
276 Slater, John C., Introduction to Chemical Physics (McGraw-Hill, 1933), 374. 277 Heitler, Interview with Heilbron, March 18, 1963, 18. Emphasis added. 278 Woody, “Putting Quantum Mechanics to Work in Chemistry”, on S612-S627 and S614-15.
98
small molecules the CI method can still grow complex, and even with modern
computation, only small systems can be treated rigorously.279
While Heitler and London’s work has been regarded as the first discovery in
quantum chemistry, Andreas Karachalios suggests that it was the physical rather than
chemical community that responded most quickly to their work. In their homeland, “the
relationships between the pioneers of the new discipline and the German chemical
community proceeded with great difficulty.”280 Because of the weak institutional ties
between chemistry and physics at the time, many German chemists were unprepared for
the possible applications of quantum mechanics to their field. Even as late as 1928 and
1930, the meetings of the Deutsche Bunsen Gesellschaft dedicated to chemical bonding
and spectroscopy had very little impact on chemists in the country. The difficulties in
solving the wave equation did not help its reception in chemistry. It was, in general,
theoretical and experimental physicists, “many of whom had no deep knowledge of
problems in chemistry”, who laid the foundations for quantum chemistry.281
Physicists were particularly enthusiastic about how far their field of research
could expand, now that problems in chemistry had become available to them. Because of
the impression that chemical processes were fundamentally defined by a purely physical
phenomenon, the Heitler-London treatment soon became the foundation of reductionism.
In the physics community of the late 1920s and 1930s, suggestions of a disciplinary
279 See Woody, “Putting Quantum Mechanics to Work in Chemistry” and Jenkins, Zach, “Do You Need to Believe in Orbitals to Use Them?: Realism and the Autonomy of Chemistry”, Philosophy of Science, 70 (2003), 1052-1062. 280 Andreas Karachalios, “On the Making of Quantum Chemistry in Germany”, Studies in the History and Philosophy of Modern Physics, 31 (2000), 443-510, on 494. 281 Karachalios, “On the Making of Quantum Chemistry in Germany”, 495.
99
connection between physics and chemistry took on hierarchical overtones, as the wave-
mechanical valence bond tipped the scales in favour of a physical worldview. Two years
before the formulation of wave mechanics, German physicist Max Born commented how
the facts of chemistry under the framework of the quantum atom showed that the “unity
of all physical and chemical forces and their reduction to reactions between the
elementary components of matter”.282 Born believed that the “vast territory of chemistry”
was important for physicists to explore, and that this work was necessary in order for
physicists to “impose her laws upon her sister science.”283
London in particular saw the potential for an overarching program to develop the
laws of chemical valency with the principles of quantum mechanics. With the ability to
formally derive these principles through the laws of physics, they believed physics could
now “eat chemistry with a spoon.”284 Perhaps the most often quoted statement of this
kind was Dirac’s:
The underlying physical laws necessary for the mathematical theory of a large part of physics and the whole of chemistry are thus completely known, and the difficulty is only that the exact application of these laws leads to equations much too complicated to be soluble.285
For physicists who were closest to the development of quantum mechanics, the new
theory seemed to offer the promise of explaining the laws of physics and, following the
Heitler-London treatment, the laws of chemistry as well.
While it may have seemed to physicists that chemistry was theirs to explore, it
took longer to generate a response from chemists. The community that had comfortably
282 Max Born, The Constitution of Matter (London: Metheun, 1923), 52. 283 Born, The Constitution of Matter, 78. 284 London, quoted in Gavroglu and Simões, “The Americans, the Germans, and the beginnings of quantum chemistry: the confluence of diverging traditions”. 285 P.A.M. Dirac, “Quantum Mechanics of Many-Electron Systems”, Proceedings of the Royal Society of
London, Series A, 123(1929), 714-733, on 714.
100
adopted Lewis’ model, despite its physical inconsistencies, had not yet fully appreciated
the relevance of the treatment for chemistry.286 One reason for this was that chemists had
had comparatively little exposure to quantum theory and quantum mechanics. An early
publication that was aimed at giving chemists this necessary background was Nevil
Sidgwick’s The Electronic Theory of Valency.287 As Lewis had done in Valence,
Sidgwick attempted to synthesise the body of physical and chemical knowledge into a
modern understanding of valency for chemists, and emphasized the value of atomic
models in interpreting chemical facts. He stressed the importance of determining how
valences were related to the spatial arrangement of atoms, but believed that this would
come from an understanding of “the nature of valency from the chemical evidence.”288
By this time, “many chemists had already become aware of the amazing explanatory
power of the new quantum mechanics, yet it was difficult to see how this newly
developing explanatory framework would be assimilated into the chemists’ culture.”289
Sidgwick believed presenting the chemical value of a physical theory would help guide
them into learning it.
Some chemists, though, still believed that physicists had given up the most
disciplinary ground in the previous decades in terms of discoveries about the atom. In a
review of valence theory following the early dissemination of the Heitler-London
treatment, chemist Worth Rodebush suggested that from the perspective of his discipline,
Lewis’ theory was still alive, well, and useful. Physical criticisms had been levelled at
286 Heitler, Interview with Heilbron, March 18, 1963, 21. 287 Nevil V. Sidgwick, The Electronic Theory of Valency, (Oxford: Oxford University Press, 1927). 288 Sidgwick, The Electronic Theory of Valency, 51. 289 Kostas Gavroglu and Ana Simões, “One Face or Many: The Role of Textbooks in Building the New Discipline of Quantum Chemistry”, in Lundgren and Bensaude-Vincent, Bernadette (eds.), Communicating Chemistry: Textbooks and their Audiences 1789-1939 (Canton, Mass.: Science History Publications/USA, 2000), 415-449, on 418.
101
the static electron, but Lewis’ model was still valuable because it offered chemists the
rule of eight:
…a rule which works for the hundreds of thousands of compounds with so few exceptions as the rule of eight is not to be discarded lightly and the chemist who has observed the variation of properties through a series of similar molecules will be loathe to discard the idea of differences in stability. For the present at least the chemist will be inclined to use the rule of eight for what it is worth.290
Rodebush was tentative about blending chemistry and physics into a single worldview.
He believed, like Lewis and Thomson, that now “physicist and chemist are observing the
same atom”, but it was not the place of physics to tell chemists how to investigate it.291
The Lewis model was successful because it provided chemists with a way to study atoms
that suited their disciplinary needs.
Chemist C.H. Douglas Clark also emphasised in a 1928 review of valency that the
greatest success of physics in recent years had been in connecting spectroscopic evidence
to the subatomic structure of the Bohr atom.292 He was more enthusiastic than Rodebush
about the explanatory potential of theoretical physics, but similarly optimistic about its
effect on future chemical work.
The chemist has also provided, in a wealth of detail, description of a complex group of phenomena… connected with the behavior of atoms in relation to each other and in actual combination. Modern physics is making important contributions to the understanding of many problems raised in this way, and perhaps offers the surest prospect of further advance.293
Douglas Clark and Rodebush both believed that while physicists had in many ways
appropriated the study of atoms and molecules, they had appropriated it from chemistry.
290 Rodebush, “The Electron Theory of Valence”, 520. 291 Rodebush, “The Electron Theory of Valence”, 531. 292 Douglas Clark, “Some Physical Aspects of Atomic Linkages”, 232, Rodebush, “The Electron Theory of Valence”, 511. 293 Douglas Clark, “Some Physical Aspects of Atomic Linkages”, 269.
102
Quantum physics was successful because it had confirmed the laws of periodicity and
classification that chemists had already set down, meaning physics had learned more
from chemistry than chemistry had from physics.294
Douglas Clark, Rodebush and Sidgwick recognized the theoretical significance of
the quantum-mechanical atom for chemistry. Their optimistic outlook about these
developments was not motivated by a hope for theoretical reduction, but by theoretical
convergence. To these men, wave-mechanics and the Heitler-London treatment supported
an autonomous chemical perspective on the valence bond. It was now “wonderfully
significant” that so many physical discoveries had confirmed chemical results like the
classification of the periodic table, but “a long road” still lay ahead before physicists
could contribute a full understanding of the “complex group of phenomena, remarkable
alike in its manifold variety and often in its lack of agreement with anticipation.”295 Even
as late as 1937, textbooks could still teach Lewis’ version as “the most useful picture of
the molecule for the purpose of interpreting chemical phenomena.”296 As wave
mechanics confirmed the chemical facts that were already known, a belief persisted that
chemical observations of the late 1920s would be “given more consideration in theories
of atomic structure than they [had] been in the past.”297
The true impact of the Heitler and London treatment on chemistry, as far as the
chemical community was concerned, was in providing a path for future research where
theoretical physics could go hand-in-hand with the progress of chemistry. Interest in the
new physical treatment of chemical bonding arose as an extension of the allegiance to the
294 Rodebush, “The Electron Theory of Valence”, 511. 295 Douglas Clark, “Some Physical Aspects of Atomic Linkages”, 232 and 269. 296 Park, “Chemical Translators”, 33-4, quoting Herbert B. Watson in Modern Theories of Organic Chemistry (1937), 14. 297 Rodebush, “The Electron Theory of Valence”, 511.
103
Lewis atom that was already in place. Rather than take away control of the chemical bond
from chemists, Heitler and London’s work suggested chemistry and physics could at last
meet at a common point of research. However, wave mechanics had changed the nature
of chemical bond research such that it would have to be practiced in a new physico-
chemical domain: chemical physics.
The remainder of this chapter will deal with how this disciplinary potential was
realized in the valence bond research of the American chemist Linus Pauling. Pauling
began his career as a physical chemist with the goal of studying the chemical bond, and
during his postdoctoral research from 1926-7 was one of the first Americans to study
wave mechanics from its German masters. Over the late 1920s and early 1930s, Pauling
pursued a program of research to directly extend the Heitler and London treatment into a
system of structural chemistry consistent with Lewis’ chemical bond. Pauling’s study, in
which he supplemented a chemically autonomous perspective on the atom and bonding
with the tools of theoretical physics, provides evidence of an emerging, new physico-
chemical worldview.
3. Pauling’s early career
3.1 The ‘boy professor’ meets physical chemistry.
Linus Carl Pauling was born on February 28, 1901, in Portland, Oregon.298
Pauling’s father Hermann, a pharmacist whose job required him to keep long hours and
move often during his children’s early years, died as the result of a severe stomach ulcer
298 Biographical information on Pauling can be found in Jack D. Dunitz, “Linus Carl Pauling: 20 February 1901 – 19 August 1994”, Biographical Memoirs of the Fellows of the Royal Society, 42 (1996), 316-338, and Thomas Hager, Force of Nature: The Life of Linus Pauling (New York: Simon & Schuster, 1995), and Ted Goertzel and Ben Goertzel, Linus Pauling: A Life in Science and Politics (New York, NY: BasicBooks, Harper Collins, 1995).
104
when Pauling was ten years old. With the family’s savings, his mother, Belle, bought a
boarding house in Portland, where Pauling and his sisters Pauline and Lucille grew up.
Family life was strained during these years, and young Linus withdrew to hobbies like
reading and rock collecting. It was during this time that he discovered chemistry, after
being introduced to a friend’s chemistry set. Eager to learn, young Pauling transformed
an area of the boarding-house basement into a home laboratory, and chemistry began to
fill a need for him to find something at which he could excel.
Pauling’s university career began with a degree in chemical engineering at nearby
Oregon Agricultural College (or OAC, now Oregon State University) in Corvallis, an
institution created in the wave of late-nineteenth century land grants by the American
government. Having had a particularly influential physics class in his last year of high
school, he was already looking forward to a degree in science before his high school
career was complete. 299, 300 At OAC Pauling moved quickly to the head of his class,
being recognized by his professors as a “most unusual student”.301 “He has the scientific
attitude”, wrote one of his professors in a letter of recommendation; “[h]e does not expect
results without hard work, but seems to delight in digging hard.”302 In his junior year
Pauling was offered a rare opportunity for an undergraduate, to teach introductory
299 Linus Pauling, Interview with John L. Heilbron, March 27, 1964, 10. LP Biographical, Box 5.020, Folder 20.8, LPP 300 By age sixteen, Pauling was ready to move ahead to university early, except for two missing credits in history. The school principal refused to waive the requirements to let him graduate early, and Pauling in turn refused to take the missing credits if it meant staying for another full year of high school. He then decided to take as much mathematics and science courses as he could in his final semester, and enrolled at OAC with enough background to take a science degree, but without a high school diploma. Hager, Force of Nature, 47-8. 301 Letter from G.A. Covell, Dean of School of Engineering, OAC, to Richard F. Scholz, Reed College, November 7, 1921. Courtesy Ava Helen and Linus Pauling Papers, Oregon State University Libraries (hereinafter LPP), LP Biographical: Academia, Box 1.002. 302 Letter from G.R. Varney, Assistant Professor of Public Speaking, OAC, to Richard F. Scholz, November 3, 1921. LP Biographical: Academia, Box 1.002, LPP.
105
courses in the chemistry department, earning himself the reputation of “boy professor.”303
Upon graduation, declaring “I shall devote my life to science”, Pauling moved on to a
graduate degree, and the beginning of a career in theoretical chemistry, at the California
Institute of Technology.304
Caltech was first founded as a preparatory school, Throop Polytechnic Institute, in
1891.305 In 1907 astronomer George Ellery Hale joined the board of trustees, and saw an
opportunity to create a research institute in the same vein as Noyes’ group at M.I.T.
Indeed, Hale’s first coup as founder was in hiring Noyes to lead the department of
Chemistry and Chemical Engineering. Still director of the Research Laboratory of
Physical Chemistry at M.I.T. at the time, Noyes agreed to take up directorship on a part-
time basis in 1915, before finally moving to Pasadena to take on the position full-time in
1919.306 In the intervening years, he brought to the school some of the country’s most
promising young chemists by sending his M.I.T. students west. By the time he arrived in
California in 1919 he had prepared for himself a “fully staffed research laboratory”.307
After Hale successfully lobbied for donations to support the new Institute, Robert
Millikan found enough incentive to join as well.308 Buoyed by the combined leadership
of these three men, Caltech quickly became an important centre for American science,
303 Hager, Force of Nature, 62-4, Pauling, Interview with Heilbron, March 27, 1964, 7-10. One of his more notable students at this time was Pauling’s future wife, Ava Helen, who was a student in his chemistry course for home economics majors. 304 Letter from Pauling to Edwin T. Reed, College Editor, OAC, July 17, 1921. LP Biographical: Academia, Box 1.002, LPP. 305 For a general history of Caltech and its development as an elite research institution, see Judith R. Goodstein, Millikan’s school: a history of the California Institute of Technology (W.W. Norton, 1991). 306 The move was somewhat bittersweet for Noyes, as by this time M.I.T. was feeling more pressure to change to a mandate more fitting engineering and applied sciences programs than a teaching and research-based institution of pure science. Face with the choice between retaining Noyes and a chemical engineer, William H. Walker, M.I.T. asked Noyes to resign. Feeling his best days were behind him, his part-time absences at Caltech made the decision easier for Noyes. Servos, Physical Chemistry from Ostwald to Pauling, 262. 307 Geiger, To Advance Knowledge, 184. 308 Kevles, The Physicists, 155-7.
106
that embodied a spirit of research beyond the simple search for practical applications of
science.
Noyes’ most profound influence at Caltech was, as it had been at M.I.T., as a
teacher of chemistry. By the time Pauling inquired about the graduate program there in
1922, Noyes’ field of vision had extended beyond the program initially set by the Ionists.
His Chemical Principles had by this point set a standard in the field for the mathematical
knowledge now required for chemical research.309 With a fresh start at Caltech his
departmental research mandate shifted from aqueous solutions to “potentially productive
areas of research in other fields” than Ostwald’s original programme.310 Prominently
among these was the new field of x-ray crystallography, for which the father and son
founders, British physicist William Henry (W.H.) and William Lawrence (W.L.) Bragg,
were awarded the Nobel Prize in physics in 1915. Like Lewis’ department at Berkeley,
Caltech offered a place where physical chemistry could be practiced on its own terms as
part of a maturing tradition of American science, rather than an emulation of a German
model.311
While Pauling had also applied to Harvard, the University of Illinois and
Berkeley, Caltech offered the best combination of a strong program and financial
assistance, which he certainly needed. Considering Lewis’ role there, Berkeley had been
a “particularly attractive” choice for Pauling, but when he heard no response to his
309 Robert S. Mulliken, Oral History Interview, February 1, 1964, Niels Bohr Library, American Institute of Physics, College Park, MD, (hereinafter AIP), and other repositories of the Archive for History of Quantum Physics. 310 Ernest Smith, Memorial for A.A. Noyes (Untitled), 5. LP Correspondence Folder 278.6, “Materials gathered for Pauling’s article on A.A. Noyes for The Dictionary of American Biography.” LPP. 311 Robert E. Kohler, “The Ph.D. Machine: Building on the Collegiate Base”, Isis 81 (1990), 638-662, on 658-60.
107
application, his interests shifted to Caltech.312, 313 As he wrote to Noyes, he felt he had to
emphasize that research “would not be strange” to him because of his experience
teaching and working as a student assistant, and told the Caltech chemist “I feel confident
I can become a really good worker in Physical Chemistry.”314 A week later, “very
favourably impressed” with Pauling’s qualifications and his interest in pursuing “purely
scientific research”, Noyes offered him a graduate assistantship in chemistry at
$750/year, telling Pauling he looked forward “with the greatest pleasure” to welcoming
him into Caltech’s “little family of research men”.315
Although Pauling had become a proficient chemist during his time at OAC, he
found he needed a great deal of preparation to augment his undergraduate studies before
he could make the transition to the new physical chemistry that Noyes was promoting.
On Noyes’ advice, during the spring and summer of 1922 he studied the Braggs’
foundational crystallography text X-rays and Crystal Structure (1915), learned the
elements of crystallography with his professors at OAC, worked through the general
course set in Noyes’ Chemical Principles, and put in “quite a bit of time” studying
French and German.316 Pauling found himself particularly weak in mathematics and
higher physics, mainly since there had been a “long period” in his undergraduate studies
where he did not have an opportunity to study mathematics, and “didn’t know enough” to
study it on his own.317 The nature of theoretical physics of the time made it especially
312 Hager, Force of Nature, 71. 313 He later learned that his application had gone directly into the “reject pile”, because Lewis simply didn’t recognize Oregon Agricultural College. Linus Pauling, Interview with John L. Greenberg, California Institute of Technology Archives, May 10, 1984, 4-5. LP Biographical Box 5.030, Folder 30.1, LPP. 314 Letter from Pauling to Noyes, February 4, 1922. LP Correspondence, Folder 278.1: Noyes A.A. LPP. 315 Letter from Noyes to Pauling, February 11, 1922. LP Correspondence, Folder 278.1: Noyes A.A. LPP. 316 Letter from Pauling to Noyes, February 20, 1922. LP Correspondence Folder 278.1: Noyes, A.A. LPP. To obtain a copy of the Braggs’ text, Pauling wrote to the state library in Salem. 317 Pauling, Interview with John L. Heilbron, March 27, 1964, 6-7.
108
important for him to correct this. By the time he arrived at Caltech, the courses he took
for his degree in physical chemistry were essentially courses in physics, with his single
course in chemistry being chemical thermodynamics with Noyes.318
In later recollections, Pauling remembered a feeling that there was no other place
where he could have had better training for his career than in Pasadena, reflecting Noyes’
goals of creating a progressive research environment. “The calibre of the professors was
so high,” he recalled in an interview, “the place was small, the number of graduate
students was small. There was freedom from bias determined by an old past history, all of
this made it a great place to work.”319 Although he felt it “wouldn’t have been any
mistake to go to Berkeley”, Noyes’ department offered an important blend of not only
physics and chemistry but of the older Ionist programme and the new physical chemistry
of 1922.320 Noyes saw Pauling as something of an early test case for new directions of
research: he identified his proficiency early on, and wanted to see what he would be
capable of doing. The training he would receive at Caltech and the timing of his degree
gave Pauling an exceptional opportunity to study physical chemistry.
3.2 Atoms in the grasp: physical chemistry meets the new physics
At the time Pauling began his graduate degree, physicists and chemists in the
United States were coming to terms with the fact that studying atoms was becoming a
more and more viable prospect. Over the 1920s, American physicists and, to a lesser
extent, chemists, slowly integrated the results of first quantum theory, then quantum
mechanics, into graduate instruction and their own research. This coincided with an 318 Hager, Force of Nature, 75-80. 319 Pauling, Interview with Greenberg, May 10, 1984, 58. 320 Pauling, Interview with Heilbron, March 27, 1964, 11.
109
expansion of American academic institutions, and expansion of theoretical research
programs, continuing from the development experienced in the early twentieth century.
From the early 1920s onward, senior physicists in the United States “made the
acquisition of theorists a major part of their expansion programs”, either from their own
country or from Europe, in an effort to keep up with the theoretical developments taking
place in Europe.321 An experimentalist tradition still prevailed, and many older physics
faculty were “experimentalists who had very little use for theory”.322 This trend was
challenging to overcome. The physicist John Hasbrouk Van Vleck, who was the first to
defend a dissertation in theoretical physics at Harvard, under Edwin C. Kemble, believed
that if the American tradition in theory had been “the equal of what it was in
experiment”, theoretical science “would have come of age earlier” in the country. 323, 324
As the quantum revolution came mid-decade, Americans had to build theoretical
traditions almost from scratch at the same time they were trying to keep up with quantum
physics. With few active centres where the newest theory was being practiced (Harvard,
Princeton, Caltech and Berkeley among them), it was hard for Americans not to feel a
sense of inferiority toward their European counterparts, while they were “just beginning
to catch up” on what those overseas were already expert in.325
In general, it was the youngest researchers at these localized centres of
progressive research, such as Harvard and Caltech, who became most adept at working
with quantum theory. Although these men had fewer role models to look to than the older 321 Kevles, Daniel, The Physicists (fourth edition) (Harvard University Press, 1995), 168. 322 John H. Van Vleck, “Biographical Information”, AIP, 33-4. 323 Van Vleck, “Biographical Information”, 5. Van Vleck recalled that in the early 1920s, there was still some question “as to whether a purely theoretical thesis for the doctorate would be accepted by the Harvard physics department”. 324 Van Vleck, “Biographical Information”, 11. 325 S.S. Schweber, “The Young John Clarke Slater,” Historical Studies in the Physical Sciences, 20 (1990), 339-406, 356, quoting Philip Morse.
110
generation of experimentalists, they did not experience the lag in literature that
nineteenth-century Americans did. John Clarke Slater and van Vleck recalled that in their
early years at Harvard that their small group kept up rapidly with Bohr’s papers and each
new edition of Sommerfeld’s Atombau und Spektrallinien as it came out, and regularly
read the Zeitschrift für Physik.326 Van Vleck believed it was only the younger physicists
who were “ever able to get quantum mechanics in their bones” and even mathematicians
had a harder time grasping the important concepts than some physicists did.327 Once the
new generation began actively researching the new theoretical discoveries and more
young men entered the field, the potential for American physics grew almost overnight,
leaving the scientists who lived through the early days of quantum mechanics feeling as
though they had “lived through an orgy.”328
Along with the new trend toward theoretical physics came fresh interest in the
problem of the atom. The generation that learned quantum physics with ease was also the
first generation that had not been taught to be sceptical about the atomic theory. Before
Slater began his graduate work, he had thought of atoms simply as “something that [he
had] read about in the Phil Mag”, but when he arrived at Harvard he found they were
being treated as “natural things to think about” and study.329 The “kind of prejudice that
Ostwald had” about atoms had vanished from Harvard by that time, and “everybody was
ready” for the changes that were about to come.330 Pauling’s first exposure to the
chemical bond had come in a similar context, while he prepared for his lectures at OAC
in 1919, and stumbled across Langmuir’s early papers while researching material for his
326 John C. Slater, Oral History Interview, October 3 and 8, 1963, AIP, 14. 327 John H. Van Vleck, Oral History Interview, October 2, 1963 (Session Two), AIP, 4. 328 Van Vleck, “Biographical Information”, 9. 329 Slater, Oral History Interview, 11. 330 Slater, Oral History Interview, 21.
111
lectures.331 The prospect of understanding chemical reactions on a foundational, atomic
level left Pauling impressed with what modern physical chemistry might be capable of.
Augmenting the theoretical developments of 1920s physics was the rise of the
experimental method discovered by the Braggs in the previous decade, to determine the
structure of crystals using x-rays. This work was based on Max von Laue’s original
discovery (which had earned him the same prize in 1914) that x-rays were a form of light
and could be diffracted by the regularly-arranged atoms of a crystal form.332 When a
homogenous or “white” x-ray beam was passed through a crystal it created a symmetric
pattern of spots on a photographic plate. Analysing Laue’s work led Lawrence Bragg to
derive a relationship, now known as the Bragg Law. This law stated a correlation
between the angle of incidence (θ) and wavelength (λ) of the diffracted rays, and the
distance (d) between the atomic centres in a crystal plane:333
nλ = 2d sinθ, for n a whole number.
From 1912 to 1915, the father and son collaborated on an experimental method to analyse
the structure of crystals using this principle.
The reflection method, as the Braggs’ called it, employed an instrument called an
x-ray or ionization spectrometer to direct a heterogeneous beam of x-rays toward one
331 Pauling, Interview with Heilbron, March 27, 1964, 8. 332 Von Laue’s original results, obtained with assistance W. Friedrich and P. Knipping, can be found in M. Von Laue, W. Friedrich, and P. Knipping, “Interferenz-Erscheinungen bei Rontgenstrahlen,” Sitzungsberichte der Bayerische Akademie der Wissenschaften, 42 (1912):303-322. For a brief history of the Braggs’ discovery of x-ray crystallography, and von Laue’s foundational diffraction experiment, see P.P. Ewald, (ed.), Fifty Years of X-Ray Diffraction (International Union of Crystallography: Utrecht, 1962) and Gasman, L, “Myths and X-rays,” The British Journal for the Philosophy of Science, 26 (1975), 51-60. For more on the transfer of the experimental method to the United States through Noyes, see also Servos, Physical Chemistry from Ostwald to Pauling, 136, 272-4. 333 Lawrence Bragg’s extension of von Laue’s results and the first analyses of crystals with x-rays can be found in W.L. Bragg, “The Diffraction of Short Electromagnetic Waves by a Crystal,” Proceedings of the Cambridge Philosophical Society, 27 (1912),43, and W.L. Bragg, “The Structure of Some Crystals as Indicated by their Diffraction of X-rays,” Proceedings of the Royal Society of London, Series A, 89 (1913), 248.
112
face of a crystal, and recorded the angle of reflection using an ionization chamber which
detected the reflected rays.334 After securing the angle of reflection (θ) and wavelength of
the rays (λ) with this method, one could obtain a measurement for the distance between
the planes of atoms parallel to each crystal face, using the Bragg Law. Since each type of
crystal had several unique faces, each corresponding to a different planar orientation of
atoms, a series of reflection measurements on each crystal face would give an
understanding of the spacing between atoms. The Braggs’ method revolutionized
crystallography, allowing physicists and chemists to ‘see’ inside crystals, rather than
studying only the external arrangement of their plane faces.
While Noyes did not perform any crystallography himself, he had realized the
importance of the Braggs’ method during his later years at M.I.T. Over time he sent two
of his students, Charles Laylor Burdick and Roscoe Gilkey Dickinson, to learn the
reflection method so it could be developed in his laboratories.335 During Pauling’s
graduate tenure there, Caltech had become “essentially the only place in the United States
where x-ray crystallography was being pursued.”336
By the mid-1920s, despite the challenges of the quantum theory and even greater
challenges of quantum mechanics, the atom had risen to the forefront of physical and
physical chemical research. In the wave of promise that quantum mechanics brought, the 334 An extensive discussion of the role of the Braggs’ x-ray spectrometer can be found in W.H. Bragg, and W.L. Bragg, X-Rays and Crystal Structure (London: G. Bell and Sons, 1915). While the spectrometer was largely W.H. Bragg’s work, and the Bragg Law due to his son, the community attributed both developments to him. He was anxious to rectify this assumption in the preface to X-Rays and Crystal Structure, noting “my son is responsible for the ‘reflection’ idea which has made it possible to advance, as well as for much the greater portion of the work of unravelling crystal structure to which the advance has led.” vii. 335 Servos, Physical Chemistry from Ostwald to Pauling, 136; Author Unknown, “Research Projects in Chemistry: Physical Chemistry”, LP Safe Folder 2.017, LPP; Letter from Pauling to W.L. Bragg, September 16, 1929, LP Correspondence Folder 30.1: Bragg, W.L., LPP. Pauling’s letter to Bragg indicates Caltech had in its early years an ionization spectrometer like the Braggs’, but that it was dismantled some time before 1929. 336 Pauling, Interview with Greenberg, May 10, 1984, 16.
113
weaknesses of both the Bohr and Lewis models now seemed less damaging to the cause
of atomic research, because there was now a clear path toward fixing them. While it was
“obvious” to physicists that they “couldn’t do anything in the form of a static atom” like
Lewis’, it was seen as indication of “some three-dimensional picture which Bohr had not
got at that period.”337 And while the rigid disciplinary boundaries between European
physicists and chemists prevented collaboration between them, the timely expansion of
American theoretical physics and physical chemistry allowed for men from both fields to
take an interest in atomic theory. The disciplinary boundaries were particularly permeable
at Caltech, where physicists and chemists regularly attended each other’s seminars and
formed a very amorphous community of researchers.
Pauling’s first forays into atomic researches were applications of the Braggs’
technique to unsolved crystal structures which, in the early 1920s, were numerous and
complex to solve, even for simple cases. His first project was a study of molybdenum, in
collaboration with Dickinson, after which point he never looked back. He began
“tabulating all of the structure determinations” that had been made with x-ray analysis, to
try to find how far apart atoms were from each other in each crystal, to bring atomic size
“into some kind of coherence with the periodic table.”338 From 1922 to 1925, Pauling
completed a series of papers that culminated in his thesis, a series of determinations of
crystal structure. In these years he built up a body of knowledge about the sizes of atoms,
ions, and atomic radii from crystal data, which he believed would contribute a more
coherent understanding of structural chemistry.
337 Slater, Oral History Interview, on 24 and 2. 338 Pauling, Interview with Greenberg, May 10, 1984, 17-18, Pauling, Interview with Heilbron, March 27, 1964, 11.
114
As he came to his final year of graduate work, Pauling had become an
experienced x-ray crystallographer and a
proficient in theoretical physics. Having had
little formative training in physics before
arriving at Caltech, his graduate education gave
him a timely education in quantum theory. His
doctoral work was essentially an immersion in
theoretical physics, which allowed him to attack
“theoretical problem of whatever sort came
along.”339 Like Slater and van Vleck’s
experiences at Harvard, Pauling’s work
emphasised the most pressing problems of the
time: general problems of quantum theory, and
modelling the atom on physical terms. As he
completed a degree in physical chemistry, Pauling had training in modern theoretical
physics almost the equal of any American doctoral candidate of theoretical physics in
1926. His interest, as a chemist, was in studying the structural properties of the atom, and
the best way to do this was to turn to the methods of physics.
Figure 1: Pauling dressed for his PhD Convocation at Caltech, 1925. Used with permission of Oregon State University Special Collections.
3.3 The chemical bond leads to physics: Pauling as theoretical physicist
As he looked toward post-doctoral research, Pauling chose to apply this
background in physical chemistry to the problem of chemical bonding. This required an
intensive program of study in the physics of the atom, at a higher level than what he had 339 Pauling, Interview with Heilbron, March 27, 1964, 16.
115
already learned at Caltech. Although American science had moved to a point where it
was not necessary for Americans to take their degrees in Europe to perform research at a
sophisticated level, Germany was nonetheless the major centre of research for theoretical
physics in the 1920s. Pauling had considered studying there for some time, with Noyes’
encouragement, and the “excitement generated by the discovery of quantum mechanics
accelerated” his decision.340
An application to the International Education Board for research funding indicates
Pauling’s intention was to use the methods of modern physics to make an intensive study
of Lewis’ non-polar bond. 341 While Lewis had connected ionic and covalent types under
the framework of the cubic atom, valence bond theory still lacked an explanation for
electron-pairing: this same gap in the understanding of bonding was what led Heitler and
London to their discovery in 1927. When Pauling planned his post-graduate work, wave
mechanics had not yet arrived, and the Bohr theory gave the only theoretical tools
available for studying the physical properties of electrons. His letters and documentation
of this time are further evidence that his interest in the problem predated the discovery of
quantum mechanics. Pauling proposed to work primarily under Sommerfeld, at the
Institute for Theoretical Physics in Munich, during the summer and winter of 1926.
Outlining his plan of study in a letter, Pauling told Sommerfeld that a “large number of
chemical facts” had led him to the belief that “the non-polar bond consists of two
340 Robert Paradowski, The Structural Chemistry of Linus Pauling (Ph.D. Thesis, University of Wisconsin, 1972), 420. 341 Linus Pauling, “Fellowship Application to the International Education Board, 1926”, LP Biographical: Academia Box 1.008, “California Institute of Technology: Graduate Studies – LP Student Transcripts, Class Notes, etc., 1922-1925, No Date”. LPP.
116
electrons in motion in orbits about two atomic nuclei”, and he wished to “make a study of
some configuration of this type.”342
Much of these ideas on the chemical bond had already emerged during Pauling’s
final year of doctoral work at Caltech, when he had made a study of bonding in the
benzene molecule using the planetary model.343 The “continued success” of the Bohr
model, despite its flaws, suggested to Pauling that physicists were willing to ascribe “a
certain reality” to the concept of electrons rotating in orbits about a positive nucleus.344
He hypothesized that introducing another nucleus into the neighbourhood of an electron
orbit would increase the orbit’s size. To form a valence bond, he suggested, the orbit
would expand to encircle nuclei of both bonding atoms.
Benzene had been a structural challenge since the nineteenth century because of
difficulties in confirming a stable molecular arrangement that would account for the
properties of aromatic compounds. In 1865, Kekulé proposed that benzene’s six carbon
and six hydrogen atoms were arranged in a closed chain, or ring.345 The following year
he proposed a hexagonal arrangement, with alternating double and single bonds (with
double bonds between carbons 1-2, 3-4, 5-6, or 2-3, 4-5, 6-1, respectively) between each
carbon in the ring, and later suggested the true structure oscillated between the two
isomeric forms. There were three main rival structures for the hexagonal ring. First was
suggested by Albert Ladenburg, in which two triangular groups of three carbon atoms
were joined in a prism structure, now known as ‘prismane’. Second, James Dewar
342 Letter from Pauling to Sommerfeld, October 20, 1925 (draft), LP Safe 3.018, LPP. On the back of the draft is a rough declination of the German verb sein, indicating the final letter was likely written out in German. 343 Linus Pauling, “The Dynamic Model of the Chemical Bond and its Application to the Structure of Benzene”, Journal of the American Chemical Society 48 (1926), 1132-1143. 344 Pauling, “The Dynamic Model of the Chemical Bond”, 1132. 345 Russell, A History of Valency, 243. See Russell, Chapter 12, for a survey of the history of benzene.
117
suggested in 1867 that instead of three double bonds evenly distributed over the
molecule, there were instead two positioned at opposite edges (joining carbons at
postions 2-3, and 5-6), with the third reaching across the molecule (joining carbons at
positions 1-4). Third, in 1867, Adolf Claus proposed a centric structure in which all
bonds were directed toward the centre of the molecule like the spokes of a wheel.
Kekule’s ring structure of alternating isomers was generally accepted toward the
late nineteenth century, but brought challenges of how to model the distribution of
electrons in the ring, when no single structure could be confirmed. With the framework of
the Bohr model of the atom, Pauling proposed a planar structure in which the four
electron orbits of each carbon atom extended across their two neighbouring carbon
atoms, a single hydrogen atom, and the carbon atom in the direct opposite position in the
ring. The six orbits of this last type extended across the centre of the ring in a criss-cross
fashion.346 The unique stability of benzene prompted Pauling to postulate the existence of
“stable configurations other than the noble gases”, which was among the questions he
hoped further training in theoretical physics would help him answer.347 A few years later,
after adopting the wave mechanical treatment of valence bonding in the Nature of the
Chemical Bond, Pauling was able to demonstrate the stability of the Kekulé ring
structure.348
What is curious about this early work of Pauling’s was the nature of his
reasoning, a mixture of that of a theoretical physicist and structural chemist. This
research clearly spoke to the interests of Lewis’ program, rather than a physical study of
346 Pauling, “The Dynamic Model of the Chemical Bond”, 1139-41. 347 Pauling, “Fellowship Application to the International Education Board, 1926”, 6. 348 Linus Pauling, and G.W. Wheland, “The Nature of the Chemical Bond. V. The Quantum-Mechanical Calculation of the Resonance Energy of Benzene and Naphthalene and the Hydrocarbon Free Radicals.”, Journal of Chemical Physics 1(1933), 362-374.
118
the electron orbits and the nature of the bonds themselves, but also situated within the
framework of the physical atom. Just as Lewis had initially speculated before he
published on the cubic atom, Pauling believed that bonding under this model implied a
“sharp demarcation” between polar and non-polar types of bonding, because the Bohr
model, with its quantized orbits, did not afford any mechanism of gradual transition
between them.349 He proposed that the terms polar and non-polar be abandoned, and
replaced with “ionic” and “molecular”, to denote the transfer or sharing of electrons in
bonding, respectively. In employing the contemporary physical model of the atom, rather
than the chemical one, Pauling could make speculations about the underlying structural
properties of chemical substances, and pursue questions that Lewis had raised in his
initial theory.350 Employing the tools of the Bohr model, however, forced him to assume
limitations on transitional bonding types, because the physical model allowed the
existence of only extreme polar and non-polar cases.
Also of note about this period in Pauling’s career is that he did not collaborate
with Lewis on chemical bond research. While Noyes supported Pauling’s plan to study in
Europe, he was also well aware of his promise as a researcher, and was anxious about
interests that might lure him away from Caltech permanently. Initially, Pauling had
planned to remain in California for the spring of 1926, giving him time to continue
crystallography research at Caltech and visit with Lewis at Berkeley. Afraid of losing his
most promising student, Noyes convinced him to apply for a Guggenheim fellowship,
which would allow him to leave for Munich before the summer term started, but prevent
349 Pauling, “The Dynamic Model of the Chemical Bond”, 1136. 350 Pauling, “The Dynamic Model of the Chemical Bond”, 1133.
119
him from staying long enough to make an extended visit with Lewis.351 Noyes proved
successful in his bid, and Pauling was awarded the Guggenheim, but it was years later
before he realized that Noyes “was determined that [he] wouldn’t set foot on the Berkeley
campus.” 352, 353 He departed for Munich in March with his wife, Ava Helen, to study the
theoretical physics that would help him understand the chemical bond, without an
opportunity to discuss the same problem at length with Lewis.
In Munich, Pauling settled in well with Sommerfeld and his group. With his tall,
lanky build and enthusiastic attitude he came across as an American “cowboy of
science”.354 Ava Helen joined her husband at lectures and in daily German lessons, and
passed on to Noyes a remark from “[o]ne of the German boys” that including herself
there were “five Americans” at the Institute.355 In one of his frequent letters to Noyes
during this period, Pauling spoke admiringly of how Sommerfeld had completely
accepted wave mechanics “as a good mathematician would”.356 If there were any
difficulties in adjusting to German scientific life, it was the different style of research
from Caltech, which favoured formal presentation of research papers and private
discussions with project supervisors rather than the collaborative discussion typical of
Pasadena seminars.357 Continually impressed with the calibre of research being done in
351 Hager, Force of Nature, 104-6. 352 In changing his plans, Pauling had to turn down a National Research Council Fellowship that he already held, which would have funded his initial research plans. Letter from Pauling to NRC Research Fellowship Board, December 28, 1925, LP Safe 3.018, LPP. 353 Pauling, Interview with Heilbron, March 27, 1964, 30. 354 Hager, Force of Nature, 126. 355 She qualified it with a wry “I fear I don’t lend much to the department.” Note at end of letter from Pauling to Noyes, May 22, 1926, LP Correspondence 278.1: Noyes, A.A. Correspondence, 1924, 1925, 1938. LPP. 356 Since “all quantum theory problems are thus reduced to the boundary conditions of the old mathematical physics.” Pauling to Noyes, December 17, 1926. 357 Pauling to Noyes, May 22, 1926.
120
Munich, Pauling gained the impression that Americans still lagged behind Europe in
terms of publications.358
Under Sommerfeld, who was well-liked and highly respected as a shepherd of the
new quantum mechanics, Pauling pursued an
intense course of theoretical physics for which he
showed great interest.359 Although his research
interests remained chemical, almost all the
lectures he attended in Munich were in physics.
The atomic structure lectures being presented for
chemists seemed “rather elementary”, in contrast
to the systematic presentation of wave mechanics
that Sommerfeld was giving.360 Realizing “atomic
and molecular chemistry” would now require the
use of quantum mechanics, Pauling devoted the
summer and fall of 1926 to theoretical research of
his own. He hoped the new physics would help
him learn “something regarding the distribution of electron-orbits in atoms and
molecules.”361 By the end of the year he was satisfied in his foundational understanding
of quantum theory, and was already looking forward to bringing the new theoretical
methods back to the chemistry seminar in Pasadena.
Figure 2: Ava Helen Pauling (centre) with Fritz London (left) and Walter Heitler (right), during the Paulings’ trip to Zürich, 1926. Used with permission of Oregon State University Special Collections.
358 Letter from Pauling to Noyes, December 17, 1926. LP Correspondence 278.1: Noyes, A.A. Correspondence, 1924, 1925, 1938. LPP. 359 Heitler described Sommerfeld as “an exceedingly decent man for whom I had the highest admiration”. Heitler, Interview with Heilbron, March 18, 1963, 4. 360 Letter from Pauling to Noyes, November 22, 1926. LP Correspondence 278.1: Noyes, A.A. Correspondence, 1924, 1925, 1938. LPP. 361 Letter from Pauling to Noyes, July 12, 1926, LP Correspondence 278.1: Noyes, A.A. Correspondence, 1924, 1925, 1938. LPP.
121
Crystallography also continued to be a primary interest of Pauling during this
time, and the scientific connections he made in Europe provided him with many
opportunities he would not have been afforded in Caltech. Among these was the
discovery of a particularly “handsome and well-built” type of crystal model
manufactured by Sommerfeld’s mechanic.362 Pauling was so enthusiastic about these
models that he advised Noyes not to attempt to have them made in Pasadena but to
purchase them from the Munich shop, telling his supervisor, “I do believe strongly that
we should have a large number of them, for I think it well to continue our work in crystal
structure.” Later, on a visit to Stuttgart, he discussed his doctoral research in
crystallography with Paul Ewald, the editor of the Zeitschrift für Krystallographie, who
requested that Pauling and Noyes send them a yearly review of the crystal work at
Caltech.363 He marvelled at the models of solid crystal solutions and the hydrogen
molecule-ion on display at the Deutsches Museum, which were “kept up-to-date through
the efforts of the leading scientific and technical men”.364
The last few months of Pauling’s fellowship were marked by extensive travel
through Europe and Britain, to meet as many of the important figures of physics and
physical chemistry as he could. These meetings included an extended stay in Zürich, with
Schrödinger as well as London and Heitler (who he had already met in Munich),
Copenhagen with Bohr, and a final trip in August of 1927 to see W.L. Bragg. Pauling at
first worried so much travelling would force the progress of his research to suffer, which
362 Pauling to Noyes, May 22, 1926, and following. In later months Pauling increased the suggested order from twenty-five to forty models, because they were “not nearly as large” as what was available at Caltech”, and the Munich ones were “beautifully and accurately made”. Pauling to Noyes, November 22, 1926, for 363 Pauling to Noyes, July 12, 1926. 364 Pauling to Noyes, May 22, 1926.
122
Noyes assuaged with the reassurance that his “many visits to scientific men and
laboratories in Europe will be a permanent satisfaction and help” to his career.365
Notably, according to his recollections of the period, Pauling became good friends
with Heitler and London but, much like his missed opportunity with Lewis the previous
year, a scientific collaboration never materialized. He and Ava Helen were among those
who celebrated Heitler’s doctoral defense with champagne afterwards, and when they
rejoined him in Zürich, struck up a connection with London as well.366 The three of them
shared regular discussion of the chemical bond, thinking of the Pauli principle and the
hydrogen molecule-ion in general terms. Although Pauling “had a vague knowledge of
what Heitler and London were doing”, he did not know the details, and was “rather
surprised”367 when their 1927 paper appeared. Part of his surprise had been in realizing
that “what they had done took so seriously” something he himself had considered to be
obvious, that two wave functions could combine to form a chemical bond. Therefore
although Pauling’s relationship with Heitler and London created a circle of discussion on
a problem of similar interest, it did not appear to provide him with any special insight on
the problem in advance of the published work.
What is clear in terms of his chemical bond research is that the year in Europe had
the major effect of adding deeper theoretical sophistication to Pauling’s work. Unlike his
first paper on the structure of benzene, which lacked any mathematical treatment of the
problem, the papers that grew from his research in Munich were highly technical studies
365 Letter from Noyes to Pauling, July 8, 1927. LP Safe, 2.019, LPP. Noyes’s support extended beyond more than professional encouragement, as he also ensured the Paulings had sufficient funds to continue touring through the spring of 1927, even after his fellowship money had run out. 366 Pauling, Interview with Heilbron, March 27, 1964, 7. 367 Paradowski, The Structural Chemistry of Linus Pauling, 425, and following pages.
123
of the atomic and ionic states.368 These papers were focused applications of
Schrödinger’s wave mechanics to the physical state of the atom, which would have fit
easily in a journal of theoretical physics.369 Publishing in the Proceedings of the Royal
Society, rather than the Physical Review, offered the added benefit of a wider readership,
where his work could be seen by the broader physics community.370
The first paper, published in the Royal Society’s proceedings, was a purely
physical study of the system given by Schrödinger’s equation, in which Pauling
developed a procedure to evaluate screening constants. These were values used to find
the effective nuclear charge (Zeff) (given by empirical values) on a many-electron atom,
while using an equation “theoretically applicable only to a hydrogen-like atom.”371
Qualitatively, the screening constant was explained as an effect “due to the action of
electrons which are nearer the nucleus than the electron under consideration.” The
method Pauling proposed would allow for the evaluation of theoretical properties of
many-electron systems, such as diamagnetic susceptibility and polarisability, apart from
the spectral analyses which were then the norm. Concluding the paper, Pauling connected
this work with his x-ray crystallography data, showing that x-ray diffraction in crystals
confirmed the orientation of Schrödinger’s eigenfunctions, which gave the position and
orientation of bonds. 368 Linus Pauling, “The Theoretical Prediction of the Physical Properties of Many-Electron Atoms and Ions. Mole Refraction, Diamagnetic Susceptibility, and Extension in Space”, Proceedings of the Royal Society of London, Series A 104 (1927), 181-211; Linus Pauling, “The Sizes of Ions and the Structure of Ionic Crystals”, Journal of the American Chemical, 49 (1927), 765-791. 369 Pauling had intended to develop the first paper, “Theoretical Predition of Physical Properties of Many-Electron Atoms and Ions”, for the Physical Review, but instead sent it to the Royal Society on the advice of Sommerfeld, who communicated the paper for him as a foreign member. This decision seemed to satisfy Pauling, who had expected the Physical Review would request a number of corrections and delay publication, and who felt Sommerfeld’s kindness to him during his stay in Munich warranted him “to act in accordance with his desires.” Pauling to Noyes, December 17, 1926. 370 Van Vleck, “Biographical Information”, 4. 371 Pauling, “The Theoretical Prediction of the Physical Properties of Many-Electron Atoms and Ions”, 181, and following.
124
His second paper, on the sizes of ions and ionic crystals, dealt almost entirely with
these x-ray analyses, and demonstrated how a theoretical knowledge from the method of
determining screening constants could be guided by crystal spacing data. The second
paper was clearly presented toward a chemical audience, while the first dealt with strict
presentation of the theoretical methods. At this stage, Pauling showed a talent for the new
physical tools, but also an ability to make their value known to audiences on both sides of
the physico-chemical boundary.
It was not hard for Pauling to develop an enthusiasm for the new physics. By the
end of 1926, he had already begun to consider the “numerous” possible extensions of the
wave mechanical treatment to problems in chemistry. His second paper, on the sizes of
ions and structure of crystals, was an extension of his first treatment of multi-electron
systems to structures of direct relevance to chemistry. As early as July of 1926 Pauling
speculated about preparing “a review of the evidence regarding the dynamic model of the
non-polar bond”, inspired by the crystallography work he had seen.372 By December
1926, the possibilities of combining the new physics with a programme of structural
chemistry had formed in his mind, and he expressed interest in making it the focus of his
research. “I think it is very interesting”, he told Noyes, “that one can see the ψ-functions
of Schrödinger’s wave mechanics by means of the x-ray study of crystals.”373 He
believed that “much information regarding the chemical bond” would result from
combining his crystallography background with his new training in theoretical physics,
and expressed hope to Noyes that he would be able to conduct this work in Pasadena.
372 Pauling to Noyes, July 12, 1926. 373 Pauling to Noyes, December 17, 1926.
125
From his work in his year abroad, from 1926-7, we see evidence that Pauling’s
encounters working in the theoretical physics community contributed a new level of
sophistication to his research. The experience was a formative one: he had left Caltech as
a physical chemist interested in the problem of the chemical bond. Incorporating the
methods of wave mechanics into his work not only gave him new physical tools to help
him study the problem, it made him further identifiable as a published practitioner of
theoretical physics. Beyond this added disciplinary identifier, however, Pauling came
away from Europe with a far different set of physical tools than what he had expected he
would find when he concluded his Ph.D. in early 1926. Instead of the application of
Bohr’s theory that he might have expected would materialize, he came away with the
entirely new physical framework of wave mechanics. This put Pauling in a specialized
category in 1927, as one of the few Americans capable of performing research in
quantum mechanics, and as a chemist who could perform this research at the same level
as a physicist with the same training.
Pauling’s research programme over the following decade, a study of the nature of
the chemical bond, exemplified this fusion of disciplinary traditions. His work applied the
methods of theoretical physics (and the framework of the physical atom) to the tradition
of Lewis’ valence bond theory that preserved a chemically autonomous view of bonding.
In so doing, Pauling’s The Nature of the Chemical Bond formed an important part of the
emerging new physico-chemical framework of chemical physics and quantum chemistry.
Later in Chapter 4, we will see how his structural chemistry and applications of wave
mechanics to chemistry allowed him to form a valuable identity within the new chemical
physics community as a practitoner of the newly formed science of quantum chemistry.
126
As a discipline that brought together several previous disciplinary threads with the
inspiration of new discoveries, chemical physics was a broad collection of sub-traditions
and specializations. In the various communities where his work found relevance, Pauling
may genuinely be identified as a quantum chemist, a theoretical chemist, and a chemical
physicist. The following section will show how his early studies of the chemical bond,
from 1928-1935, combined previous chemical and physical traditions to make the
valence bond a hybrid, physico-chemical theory.
4. The physical nature of the chemical bond
4.1 Pauling’s program begins
Pauling returned to Caltech from Europe in 1927 to find an Assistant
Professorship waiting for him at Caltech, secured by Noyes in his absence. Noyes
believed the application of atomic theory to problems of structural chemistry was one of
the most important problems in chemistry, and at the same time “the most backward,
considering the little yet accomplished in relation to its tremendous possibilities”.374 He
was eager to ensure Pauling had all the resources he needed to pursue this research at
Caltech. Noyes had eagerly followed his former student’s progress over the year, and
recommended him well to the faculty.375 Although he did not “understand very well the
theoretical side of the wave-mechanics treatment”, Noyes could see Pauling had
identified some problems of “very great importance”, and he was “much impressed” with
what he had accomplished.376, 377 The faculty position, with a starting salary of $3,000
374 Letter from Noyes to Pauling, February 25, 1927, LP Safe Folder 2.019, LPP. 375 Letter from Noyes to Pauling, February 25, 1927, LP Safe 2.019, LPP, on 2. 376 Noyes to Pauling, February 25, 1927, 2. 377 Letter from Noyes to Pauling, July 8, 1927, LP Safe 2.019, LPP.
127
per year, was more than what would usually be done for someone who had completed his
doctorate so recently, but the Caltech administrators were “so confident of [Pauling’s]
future success”, in teaching and in research, that they were “very glad to do it.”378
The major point of debate over Pauling’s position was how to classify the kind of
chemistry he was practising, within the existing disciplinary categories of chemistry,
physical chemistry, and theoretical physics. For Pauling, the question of whether he
should be seen as a physicist or chemist was “more than an academic one.379 His own
impression was that despite the general opinion that the study of atoms and molecules
should be classed as physics, any study of chemical structures was chemistry, whether it
used higher mathematics or not. When Millikan objected to him teaching what was
essentially a physics course in the chemistry department, Pauling integrated the material
he had developed in Munich into a quantum mechanics course, which the administrators
“didn’t complain about”.380 It was also Noyes’ hope that he would foster ties with the
mathematicians and physicists at Caltech, to help his work in the long run. Indicating an
awareness that Pauling was working in a field that could not adequately be represented by
the name of physical chemistry,381 he was in the end given the title of Assistant Professor
of Theoretical Chemistry.382
Being one of the few experts on the new wave-mechanical theory of bonding gave
him many opportunities to lecture on it, bridging the new physics with the previous
version of chemical valency for the benefit of the chemistry community. One of the most
378 Noyes to Pauling, February 379 Pauling to Noyes, December 17, 1926. 380 Pauling, Interview with Heilbron, March 27, 1964. 381 The name had by now “come to connote those sides of chemical theory which were especially developed between 1885 and 1910.” Noyes to Pauling, February 25, 1927, 4-5. 382 Abbreviated from “Theoretical Chemistry and Mathematical Physics”. Noyes to Pauling, February 25, 1927, 4.
128
important avenues for this introduction was a series of lectures Pauling gave in the
chemistry department at Berkeley, at Lewis’ invitation. The senior chemist had visited
Caltech late in 1928 with the intention of offering Pauling a position, and while this
endeavour proved unsuccessful, a series of guest lectures was the next best offer.383 After
some haggling with Noyes to release him,384 and a series of letters and telegrams to
confirm lecture topics, Pauling agreed to offer two courses, “Quantum Mechanics with
Chemical Applications”, and “The Sizes of Ions and Other Ionic Properties.”385 These
topics reflected both the personal research interests that had brought him to the topic of
the chemical bond, and the interests of a group of chemists who had yet to learn the full
significance of the wave-mechanical formalism for their field.386
In his first few years of lecturing at Berkeley, Pauling’s lectures had the dual
purpose of presenting the theory of wave mechanics to a fresh chemical audience, and
then its value to chemistry through the Heitler-London treatment.387 The style of many of
these early lectures was not easily recognizable as chemistry, as Pauling first re-taught
the nature of atoms and molecules as systems in the new physics before addressing their
chemical properties. He then took opportunities to connect this foundational theory to the
more familiar chemical language of Lewis’ scheme. Upon introducing the symmetry
383 Pauling, Interview with Greenberg, May 10, 1984, 19. Lewis met first with Noyes to broach the subject of offering Pauling a faculty position, but Noyes had no wish to lose Pauling from his Caltech staff. On Noyes’ request Lewis did not make the offer, which Pauling did not discover until decades later. 384 Western Union Telegram from Lewis to Noyes (undated, likely December 1928), LP Correspondence 216.1 Lewis, G.N.:Reprints, Correspondence, 1913-1947. LPP. 385 Telegram from Lewis to Noyes (undated ), Western Union Telegram from Lewis to Pauling, December 24, 1928, Western Union Telegram from Lewis to Pauling, December 31, 1928, LPP LP Correspondence 216.1 Lewis, G.N.:Reprints, Correspondence, 1913-1947. LPP. 386 For a more comprehensive treatment of the early audiences for chemical physics, and the spread of textbooks like Pauling’s Nature of the Chemical Bond, please see contributions in Bernadette Bensaude-Vincent (ed.), Communicating Chemistry: Textbooks and their Audiences, especially Mary Jo Nye, “From Student to Teacher: Linus Pauling and the reformulation of the Principles of Chemistry in the 1930s”, 397-414, and Kostas Gavroglu and Ana Simões, “One Face or Many: The Role of Textbooks in Building the New Discipline of Quantum Chemistry”, 415-449. 387 Pauling, “Twentieth Lecture, April 10, 1929”, LP Speeches 1929s, LPP.
129
constraints for the wave function, where “the symmetric eigenfunction leads to molecule
formation”, and the energy producing the bond arises from the resonance energy “of the
electrons changing places,” he made clear that these electrons were the same familiar
shared pair they had been since 1916.388 This work served to bridge, for a general
audience, the known theory of chemical valency with the physical theory that had
appeared to take its place.
During his first presentations of the Heitler-London method and its applications,
Pauling worked hard to instruct Berkeley chemists on how to interpret the wealth of
physical concepts that were before them. To this fresh audience, he downplayed the
mathematical nature of his work even as he presented them with the technical details of
the resonance effect and the wave-mechanically defined electron pair. Pauling preferred
to portray himself as a chemical outsider, not a physical insider, in a world the chemical
bond had drawn him into:
I myself am no specialist in the field of quantum mechanics, for my interest is in structural problems – the structure of atoms, of molecules, of crystals, of everything – and I strive for pictures and models, taking mathematical equations with them if necessary. This viewpoint is abhorred by the quantum mechanician. So in this course we shall use what mathematics we have to use, but developing it as we go along.389
Instead of misleading chemists about the need to solve complex integrals and adopt
challenging theoretical concepts, Pauling gently introduced what they would need to
expect.
Accompanying Pauling’s exposition of the new valence bond treatment was an
enthusiasm for the physics that made it possible, which was initially coloured by some
388 Pauling, “Twentieth Lecture, April 10, 1929”, 84. 389 Pauling, First Berkeley Lecture, March 17, 1930, LP Speeches 1930s, LPP, 1.
130
shades of the reductionist attitudes shown by physicists in previous years. In the spring of
1928, Pauling had spoken eagerly about the “mighty strides” theoretical physics had
taken toward “the reduction of the phenomena included within its domain to the simplest
form”, even as it diverged from the “old road” of the Bohr theory.390 He expressed a
belief that since chemistry was “more complicated and extensive” than physics, it should
“necessarily follow physics in its development.”391 Now that chemistry appeared to
depend essentially upon the physical phenomena of resonance and electron spin, chemists
should expect to receive a thorough grounding in those portions of the new physics.
Telling his audience that the wave-mechanical treatment of valence was “more detailed
and correspondingly more powerful than the old picture” was a clear indication that
purely chemical approaches to valence would always be limited.392 If valence research
was to move forward, it would need to proceed with physical tools.
However, unlike the physicists who championed the reductionist view, Pauling’s
attitude was tempered by an equal enthusiasm for the better prospects now in sight for
structural chemistry. Problems “relating to choice among various alternative structures”
could not be solved “directly by the application of the rules resulting from the quantum
mechanics”, but the new ability to discovery physical properties of individual structures
gave chemists a powerful tool.393 In combination with the facts of chemistry and Lewis’
earlier theory, wave mechanics could move valency into a new level of sophistication. In
presenting the main points of the Heitler-London treatment in 1928, Pauling made his
390 Pauling, “The Nature of the Chemical Bond”, Draft of speech for the 1928 meeting of the AAAS, Pomona, California, June 14, 1928, p 1. LP Speeches, LPP. 391 Pauling, “The Nature of the Chemical Bond”, Draft of speech for AAAS, June 14, 1928, p 2. 392 Pauling, Linus, “The Shared Electron Bond”, Proceedings of the National Academy of Sciences, 14 (1928), 359-362, on 360. 393 Pauling, “The Shared Electron Bond”, 361.
131
chemical priorities clear by stating that the new theory was “in simple cases entirely
equivalent to G.N. Lewis’ successful theory of the electron pair.”394 Pauling not only
acknowledged the previous precedent of Lewis’ valence bond theory, but also made clear
that his own work would build on Lewis’ original statement.
While Pauling valued Langmuir and Kossel’s contributions to the octet theory, he
viewed Lewis as the father of valence theory. He believed Langmuir had made
substantial contributions to the progress of chemical bond research, and that Lewis had
been too critical of him. By contrast, however, Kossel’s work seemed to Pauling mostly
“a long discussion of electrovalence” with “nothing about covalence” and should be
ignored.395 The originality of Lewis’ work lay in the “deduction of the shared electron
pair and its properties from chemical data”, and was in Pauling’s eyes “the most striking
example of the deduction of a simple and fundamentally entirely new feature of nature
from a most complicated body of knowledge.”396 It was Lewis who had made the shared
electron pair the “basis of modern chemistry”, and he was the chemist to whom Pauling
owed the most intellectual debt.397
Lewis felt a professional connection to Pauling as one would extend the tradition
of research he had begun. “We are delighted you are going to be with us,” Lewis told him
in January of 1929, as he offered Pauling help in preparing for his stay in Berkeley.398
Pauling reciprocated the following month, telling Lewis he looked forward “with keen
394 Pauling, “The Shared Electron Bond”, 359. 395 Pauling, “G.N. Lewis and the Chemical Bond”, Notes for 1982 Meeting of the American Chemical Society, Symposium on Lewis, LP Speeches Folder 1982s.15, LPP. 396 Pauling, Linus, Fortieth Berkeley Lecture, April 26, 1929, LP Speeches Folder 1929s, LPP, on 120. 397 Pauling, “G.N. Lewis and the Nature of the Chemical Bond”, 7. 398 Letter from Lewis to Pauling, January 29, 1929, LP Correspondence 216.1 Lewis, G.N.:Reprints, Correspondence, 1913-1947. LPP.
132
anticipation” to joining the Berkeley group.399 He started the lectures “not without a
feeling of temerity” to be continuing the chemical bond story in the place it was
“invented”,400 but later grew to feel as part of the community there. As he began the 1931
lecture series he described himself not as an outsider but rather as “a member of the
group here carrying on the work begun by Professor Lewis”, fancying himself “to some
extent a student” of his even though he “never matriculated at the University of
California.”401 Pauling’s time there showed him there was “at least one stimulating place
in America” he could visit, outside of Caltech, to find valuable professional response to
his research.402 Most importantly, the chemists who heard Pauling’s lectures were
immediately presented with the chemical roots of the theory along with the
predominantly chemical context.
4.2 The nature of the chemical bond
In the series of papers, “The Nature of the Chemical Bond” which formed the
bulk of the content of the 1939 book of the same title, Pauling studied the molecular
arrangement of a wide range of compounds using this method of resonance structures.
Operating under the general principle that molecules either corresponded to a wave
function representation of a Lewis-type structure or a linear combination of resonance
structures, Pauling created a new system of structural chemistry in which the concepts of
quantum mechanics became tools that would help chemists in reasoning with problems of
molecular structure.
399 Letter from Pauling to Lewis, January 2, 1929, LP Correspondence 216.1 Lewis, G.N.:Reprints, Correspondence, 1913-1947. LPP. 400 Pauling, First Berkeley Lecture, March 23, 1931, Folder LP Speeches 19301s, LPP, 1. 401 Pauling, First Berkeley Lecture, March 23, 1931, 1. 402 Letter from Pauling to Noyes, March 21, 1929, LP Safe 2.003. LPP.
133
Beyond instructing chemists about the value of wave mechanics, the Berkeley
lectures gave Pauling the opportunity to develop his own lines of inquiry, introducing
several concepts that later grew into fundamental principles in his research. Here I regard
his work in the period from about 1928 to 1932, as an emerging synthesis of the
traditions of chemical valency and wave mechanics into his own physico-chemical hybrid
treatment of valence-bonding and structural chemistry. After using his Berkeley lectures
to develop the concepts of ionic character, transitional bond types, and resonance,
Pauling formally presented these in the 1931-1933 series of papers entitled “The Nature
of the Chemical Bond.” By the mid-1930s he had shown chemists that “many more
results of chemical significance can be obtained from the quantum-mechanical
calculations” beyond Heitler and London’s treatment, making way for “an extensive and
powerful set of rules for the electron-pair bond supplementing those of Lewis.”403 The
following section will examine how this view developed in his early studies of the
valence bond, and the phenomenon of resonance.
In Pauling’s programme, the chemical bond was much the same entity as Lewis
had postulated, but he redefined it in terms of the new physics. The eigenfunction
solutions of the wave equation corresponded to orbital types s, p, and d, of increasing
levels of energy and distance from the nucleus. The electron pair bond itself was given by
the interaction of unpaired electrons from the two bonding atoms, with opposing spin.404
Once involved in a bond, the pair could not contribute to any other bonds. Pauling’s early
presentation of this system made clear the unique consequences of the wave-mechanical
403 Pauling, Linus, “The Nature of the Chemical Bond. Application of results obtained from the quantum mechanics and from a theory of paramagnetic susceptibility to the structure of molecules.”, Journal of the American Chemical Society 53 (1931), 1367-1400, on 1367. 404 This summary taken from Pauling, “The Nature of the Chemical Bond”, JACS 53 (1931), 1369.
134
nature of bonding: given a choice between two bond eigenfunctions at the same bond
direction, the one at the smallest energy level would form the bond. Likewise, given two
eigenfunctions with the same energy level, the one with the larger value in the bond
direction would form the strongest bond. Using this set of rules, composed over his first
few years of study of valence bonding, Pauling made decisions about the structure of
previously unsolved substances, and connect a range of chemical problems to the same
framework.
The notion of valence bonds having “ionic character”, or some polar quality
despite their mainly non-polar nature (in the language of Lewis’ period), first grew from
a study Pauling made of alkali halides and halogen halides. In his 1929 lecture series, he
presented a tentative method of determining whether a bond could be categorized as
“ionic” or “shared-electron” by calculating the interatomic distance at which the energy
for the ionic state and atomic state appeared to be equal.405 From these calculations he
found that the energy of the states suggested that H2, HCl, HBr and HI were stable as
shared-electron structures, while KF, KCl, KBr and KI were ionic.
For Pauling, criteria for distinguishing between bond types came from energy: the
state that required the lowest energy would be more stable, and therefore the more likely
scenario.406 However, in some cases the energy curves obtained from calculation of
interatomic distance did not show a clear favourite. For such cases, there was “some
probability of the system taking either curve”, meaning the molecule could be stable in
405 Pauling, Linus, Twenty-seventh Berkeley Lecture, April 22, 1929, Folder LP Speeches 1929s, LPP. 406 Pauling, Twenty-eighth Berkeley Lecture, April 19, 1929, 109.
135
either the ionic or covalent state.407 Hydrogen fluoride (as we will see in the following
pages) provided one of the best examples of such a system.
In the following years, the subject of these transitional bond types took priority in
his lectures as Pauling developed his own lines of research to go beyond only exposition
of Heitler-London. In March of 1930, he introduced the topic of transitional quantum
states in his opening talk, and in February of 1932 he began by situating intermediate
bond types in the tradition previously set by Lewis.408 He suggested in 1932 that the
Berkeley community would understand the significance of his discoveries about the
hydrogen halides, because as a result of Lewis’ influence they were among the few
chemists who did not try to “classify molecules rigorously as ionic or non-ionic.”409
While in previous years, an ionic classification might have been forced on structures like
HCl or HBr, the new subtleties of the wave-mechanical treatment now allowed chemists
to regard them as predominantly covalent. Harnessing Heitler and London’s technique
now allowed him to study the continuous transition between bond states that Lewis had
postulated.
As transitional bond types became a more focused component of Pauling’s “The
Nature of the Chemical Bond” programme over the early 1930s, he showed more
conclusively that resonance and the wave-mechanical treatment provided a causal
understanding of them. If no single structure could be assigned to a molecule, it was
because the molecule was in a hybrid state, “intermediate between one extreme type and
407 Pauling, Twenty-eighth Berkeley Lecture, April 19, 1929, 109 408 Pauling, Linus, First Berkeley Lecture, March 17, 1930, LP Speeches 1930s, LPP, 1. 409 Pauling, Linus, Second Berkeley Lecture, March 3, 1932, LP Speeches 1932s, LPP, 6.
136
another.”410 Returning to his earlier studies of the hydrogen halides, he combined an
experimental analysis of bond energies with a theoretical analysis of the bond
eigenfunctions.
Pauling’s treatment set up a resonating system where the molecule was
represented as a linear combination of two possible states:
aψionic + bψelectron-pair.
This procedure was what he called a “semi-quantitative treatment” designed to approach
special cases in which a categorical assignment to either the covalent or ionic state could
not be made. In most cases, the resonance effect between the two terms was small, such
that the structure could be considered belonging predominantly to one category, but still
retaining some character of the other bond type. In extreme transitional cases, such as
HF, the energy curves favoured neither state, but using the theoretical treatment showed
it was ionic with partially covalent character. Pauling further discovered this degree of
ionic character decreased with each hydrogen halide, with HCl, HBr, and HI all showing
increasing covalent character.
Examining ionic and covalent bonds in this respect gave Pauling a further
understanding of the electronic properties of each molecule and the subtle gradations
between them. Extending the idea of ionic character, he suggested that ionic structures
contributed to many normal covalent bonds: in hydrogen, for example, wave functions
for the ionic states H:- H+ and H+ :H- contributed about 15% of the energy of the normal
molecule. However for like atoms, or for two atoms of similar powers of attraction, the
contributions for the states A:- B+ and A+ :B- “contributed equally to the bond,” leaving
410 Pauling, L. (1932a). The Nature of the Chemical Bond. III: The Transition from One Extreme Bond Type to Another. Journal of the American Chemical Society, 54, 988-1003, on 989.
137
a primarily covalent system, with the bond shared equally between the two atoms.411 In
his fourth paper in the “Nature of the Chemical Bond” series, Pauling suggested the bond
energy of a true covalent link A:B was approximately equal to half the sum of the bond
energies A:A + B:B.412 These discoveries gave him a general method of estimating the
strength of a covalent bond based on its constituent atoms.
These studies led Pauling to a further connection between extreme bond types,
with the property of electronegativity. Pauling defined electronegativity as the strength of
an atom’s ability to attract an electron to its outer shell. Differing electronegativities
between bonded atoms could influence the strength of the bond’s ionic character. A
purely covalent bond would only be created when the bonding atoms had the same
electronegativity. Using data like the additivity rule, and an examination of experimental
bond energies, Pauling was able to map electronegativities of the known elements onto
the periodic table, allowing for “the formation of a reasonably reliable scale” of attraction
for electrons on a wide scale.413 Among the properties this clarified was the divide
between metals and non-metals, where the electronegativity dividing line was zero. The
degree of variation that was apparent in ionic character showed conclusively that, through
resonance of ionic and covalent terms, a continuous transition between Lewis’ extreme
types was in fact possible.
The second major application of resonance, and the Heitler-London treatment, in
the “Nature of the Chemical Bond” series was in treating unusual and exceptional
411 Pauling, Linus, “Partial Ionic Character of Single Covalent Bonds”, Lecture 7, “The Nature of the Chemical Bond”, Lecture series at Caltech, 1935-1936, LP Biographical Academia Folder 1.011, LPP 412 Pauling, Linus, “The Nature of the Chemical Bond. IV. The Energy of Single Bonds and the Relative Electronegativity of Atoms”, Journal of the American Chemical Society, 54 (1932), 3570-3582, on 3572. 413 Pauling, Linus, “The Revised Electronegativity Scale”, Lecture 10, October 28, 1935, “The Nature of the Chemical Bond”, Lecture series at Caltech, 1935-1936, LP Biographical Academia Folder 1.011, LPP
138
structural cases. In the second paper, Pauling picked up a thread Lewis had speculated on
in Valence, of the nature of one- and three-electron bonds.414 With only one electron
available, “resonance is not expected in general” to allow for molecule formation. In
some rare cases, a stable bond could form if the right energy criterion was satisfied.
When the single electron oscillating between the two bonding atoms led to “two
conceivable states of the system with essentially the same energy”, then resonance would
create a stable system.415 While these bonds had been hypothesized to account for certain
chemical properties, such as the solubility of boron hydrides, previous methods had been
unable to confirm or deny their existence. Pauling felt this had led to “lavish use of one-
electron bonds by some English authors”, “without justification.”416
Lewis’ postulation of one-electron bonds in boron compounds led Pauling to
study boron halides and simple boron hydrides like B2H6.417 After studying energy curves
for the hydrides, with the same method as in the case of the transitional bond types, he
concluded “the one-electron bond is to be accepted for the boron hydrides, whose
existence provides the strongest evidence that the condition for the formation of this bond
is satisfied.”418 At the same time, he suggested that conditions for resonance would
probably not be satisfied unless a hydrogen atom formed part of the bond, and postulated
the existence of similar bonds in Li2+ and Na2+. Subsequent examples confirmed, with
similar methodology, the presence of three-electron bonds in the helium molecule-ion
and certain oxygen and nitrogen compounds. While he ascribed only tentative structural
414 Lewis, Valence and the Structure of Atoms and Molecules, Chapter 4, especially 97-101. Pauling, Linus, “The Nature of the Chemical Bond. II. The One-Electron Bond and the Three-Electron Bond.”, Journal of the American Chemical Society 53 (1931), 3225-3237. 415 Pauling, “The Nature of the Chemical Bond. II”, 3225. 416 Pauling, “The Nature of the Chemical Bond. II”, 3229, footnote. 417 Lewis, Valence, 98. 418 Pauling, “The Nature of the Chemical Bond. II”, 3228.
139
causes for the existence of these cases, Pauling showed that wave-mechanical methods
could bring an understanding to previously vague chemical structures.
Pauling’s most widely celebrated discovery was perhaps the existence of hybrid
orbitals, which accounted for the tetrahedral structure of the carbon atom.419 This study
followed some speculation about the changing quantization of bond orbitals. While
carbon compounds had been known to have a tetrahedral structure since the 1870s, the
new valence bond methods offered a more precise understanding of bonding electrons. In
his first “Nature of the Chemical Bond” paper Pauling showed, by comparing the
energies of carbon-hydrogen bonds that the four valence electrons in carbon were not in
the configuration 2s22p2. Their energies were intermediate between the s and p levels,
leading Pauling to conclude this “change in quantization” of orbital energy caused the
formation of four equivalent bonds, “directed toward tetrahedron corners.”420
When the time finally came in 1939 for Pauling to publish his major work, The
Nature of the Chemical Bond, the major research had already been done. All the elements
of his lectures, seminars and published papers came together in a catalogue of structural
chemistry: electronegativity, bond angles, ions in crystals, one- and three-electron bonds,
and even metallic structures were all presented in the new language of resonance. In his
examples and sketches, Pauling downplayed the mathematical formalism of the wave-
mechanical treatment in favour of emphasising the practical, experimental sources of his
conclusions, making sure they would be accessible to his chemical audience. Sprinkled
419 Pauling made this discovery simultaneously with the Harvard physicist John Clarke Slater, who also published his results in 1931. The two discoveries were regarded as independent contributions to the same problem, and did not involve a debate over priority. Slater’s role will be discussed in greater detail in Chapter 4. See also Buhm Soon Park, “The contexts of simultaneous discovery: Slater, Pauling, and the Origins of Hybridization”, Studies in the History and Philosophy of Modern Physics, 31 (2000), 451-474. 420 Pauling, “The Nature of the Chemical Bond. I”, 1374.
140
throughout with diagrams, tables and sketches of crystal units, bond angles, and ionic
radii, The Nature of the Chemical Bond was an accessible reference text for a new
generation of chemists. Though the valence bond had its roots in Lewis’ familiar theory,
it was now refined through modern physics, “dispersing the veil of mystery” in which it
had long been “shrouded.”421
4.3 The ‘physico-chemical’ nature of the valence bond
My focus on Pauling’s early research here has been to present his research in the
context of the traditions begun by Lewis and Heitler and London, to show how the
exposure to physical and chemical disciplinary influences in his early career shaped
Pauling’s chemical bond into a physico-chemical work. By proceeding from Lewis’
original statement of the chemical bond problem and carrying it out using Heitler and
London’s techniques, Pauling showed the methods of wave mechanics were an aid to the
rules of valency that went beyond what could have been accomplished with nineteenth-
century techniques. The communities in Munich, Caltech, and Berkeley accepted and
related to Pauling because they could all identify with some aspect of his research. His
Nature of the Chemical Bond was not a direct extension of the older physical chemistry,
nor was it a calculated venture into a modern chemical physics; the nature of the valence
bond in the post-wave mechanics era had simply led him to new interdisciplinary ground.
Resonance afforded Pauling, and the chemists who applied his results, a causal,
physical understanding of the structures they had previously understood in generally
qualitative chemical terms. Under his programme, the methods of wave mechanics could
421 Linus Pauling, The Nature of the Chemical Bond and the Structure of Molecules and Crystals: An Introduction to Modern Structural Chemistry (Ithaca, New York: Cornell University Press, 1939), on 2.
141
be applied nearly universally, to chemical structures of all shapes, sizes and properties.
The construction of resonance systems for these studies involved an amount of
arbitrariness, which Pauling freely admitted.422 He cautioned chemists that this weakness
should not be seen as a reason to abandon their use, since the application of resonance
structures gave chemists useful and “illuminating” structures for comparative structural
studies. In so doing, chemists could then “extend a system of structural chemistry to
include a larger class of substances than it would encompass if resonance were
ignored.”423 Whether or not the structures were real, or believable, was secondary to their
practicality and usefulness in chemical reasoning.
The advantages of incorporating physical tools into structural chemistry made it
possible for Pauling to re-define it as a physico-chemical pursuit, but nonetheless allowed
the research to retain its identity as chemistry instead of being seen as a newly annexed
physical practice. As reductionists believed, the exclusion principle and the resonance
phenomenon were accepted to be “responsible for the variation of chemical properties,
and the formation of chemical bonds in general”, but as Pauling’s studies continued he
concluded that there was “more to chemistry than an understanding of general
principles.”424 The progress made since Schrödinger and Heisenberg’s discoveries had
given chemists a new wealth of understanding about the “arrangement of atoms and the
distribution of valence bonds”, which enabled them to “begin to understand the chemical
properties of substances” better than before.425 In Pauling’s view these physical
422 Pauling, The Nature of the Chemical Bond, 11. 423 Linus Pauling, “Nature of the Chemical Bond”, Lecture 3, Caltech Lectures Fall 1935, p 10-11. LP Speeches Folder 1935s, LPP. 424 Linus Pauling, “Recent Work on the Structure of Molecules”, Manuscript for Athenaeum, California, February 7, 1935, LP Speeches Folder 1935s, on 2. 425 Pauling, “Recent Work on the Structure of Molecules,” 2.
142
discoveries, when used to study chemical structures, furthered a new chemical awareness
of the natural world.
Reviews of The Nature of the Chemical Bond (1939) were enthusiastic in their
praise of Pauling’s treatment, which was all the more successful for his chemical
audience because of his emphasis on resonance, and de-emphasis on the mathematical
nature of his discoveries. “This book,” wrote John E. Vance, “offers the chemist the first
opportunity to learn of the results of wave mechanics, as they affect the subject nearest
his professional heart, without the use of mathematics.”426 G.B. Kistiakowsky
recommended the book to “all who are not satisfied with the bond-dashes of a chemical
text and wish to get a deeper understanding of their meaning.”427 Robert Mulliken,
founder of the rival molecular orbital theory of bonding (who we will turn to in detail in
the next chapter) noted in particular of the treatment of ionic character and
electronegativity that such a subject is “particularly difficult to treat quantitatively, but
the author has made great progress by his pioneer work and stimulating new ideas”. 428
In applying resonance to a scheme of valency, Pauling drew on work from both
physical and chemical traditions. He saw Lewis as the founder of valence bond theory,
but his ability to carry out Lewis’ programme as successfully as he did relied on the new
concepts discovered in quantum physics, a connection that Mulliken also noticed in his
review of The Nature of the Chemical Bond:
[t]his is a clearly written survey of the nature of chemical bonding as seen from the viewpoint of the atomic orbital method. It is in this method that the
426 John E. Vance, “Review: Pauling, Linus, The nature of the chemical bond and the structure of molecules and crystals (Cornell University Press, 1939).”, American Journal of Science, Dec. 1939, 923-4. 427 G.B. Kistiakowsky, “Review: Pauling, Linus, The nature of the chemical bond and the structure of molecules and crystals (Cornell University Press, 1939).”, Journal of the American Chemical Society, 51 (1939), 457. 428 Robert S. Mullikan, “Review: Linus Pauling, ‘The nature of the chemical bond’”, The Journal of Physical Chemistry, 44 (1940), 827-828, on 827.
143
wave mechanics connects most easily and naturally with the traditional ideas of chemists on the nature of chemical bonds and with the Lewis theory. Hence the presentation will have wide appeal and usefulness among chemists.429
Pauling’s work gave chemists what Lewis had hoped for; a view of valence bonding that
unified the physical and chemical worldviews and appealed to the chemists’ preference
for the chemical theory of valency.
Heitler and London provided valence theory with the mechanism of the electron
pair it had lacked since Lewis’ original postulation of the valence bond, but it was not
until Pauling’s work that these two developments were synthesised into a complete
chemical theory of the valence bond. The Heitler-London treatment gave Lewis’ theory
the piece it had been missing; The Nature of the Chemical Bond made it practicable and
accessible to chemistry as a whole. Pauling’s work contributed to the growing body of
work in the borderland between chemistry and physics that had been “widely extended”
since the arrival of quantum mechanics.430 It was a part of the new chemical physics, as a
theory developed from a convergence of disciplinary traditions. After Pauling’s synthesis,
neither physics nor chemistry could claim ownership of the valence bond; it was a new
theoretical hybrid.
In Chapter 4, we will look at how this hybrid nature of the chemical bond
reflected the interdisciplinary nature of physico-chemical research in the 1920s and
1930s, by studying the work of Pauling’s contemporaries and the responses to his work in
the wider community. In particular, we will see how historical claims of reductionism
based on the Heitler-London treatment have been addressed by modern philosophers of
chemistry. The nature of the chemical bond as a physico-chemical synthesis forms both a 429 Mullikan, “Review: Linus Pauling”, 827. 430 Douglas Clark, “Some Physical Aspects of Atomic Linkages”, 232.
144
response to reductionism, but also informs deeply on the character of chemical physics
and the process of disciplinary emergence in the early twentieth century. The connections
made by Pauling and his contemporaries show that as physicists and chemists responded
to problems of common interest, they changed the shape of physico-chemical research
into a much different discipline than the Ionists had first created.
145
Chapter 4: Paths to interdisciplinarity: the chemical bond in context
“…although he calls himself a chemist, the things he actually does are almost exactly in the same line I am working in, and it seems we could cooperate very well.”431
1. Introduction
The history of the valence bond treatment from Lewis, to Heitler and London, to
Pauling, highlights a tradition of physico-chemical research created from the application
of physical methods to a foundational problem in chemistry. From the historical
perspective, the contributions of Lewis, Heitler and London, have been viewed as
groundbreaking research within chemical and physical traditions, respectively. In The
Nature of the Chemical Bond, Pauling reassembled these separate lines of research into a
physico-chemical whole, and presented a programme of structural chemistry that relied
on the tools of quantum mechanics to carry out the research problems of classical
valency. This work followed the conceptual tradition created by Lewis’ theory of
valency, applying theoretical discoveries like resonance and experimental methods like x-
ray crystallography to confirm and predict hypotheses previously made on the basis of
only chemical evidence.
The nature of valence bond theory as a synthesis of physico-chemical work
reflects the disciplinary context in which it developed, within the new chemical physics
of the 1930s. Pauling’s talents as a chemist trained in both quantum mechanics and the
technique of x-ray crystallography made him a valuable teacher and researcher in this
climate, and the career prospects offered him in his early career serve to confirm the high
interdisciplinary value of his research. Historians of chemistry have recently emphasised 431 Slater, quoted in Samuel S. Schweber, “The Young John Clarke Slater,” Historical Studies in the Physical Sciences 20 (1990), 339-406, on 401.
146
issues of interdisciplinarity through broader studies of the relationship between physics
and chemistry, in the hybrid fields of first physical chemistry in the late nineteenth and
early twentieth century, and then chemical physics in the early twentieth.432 These fields
thrived on the use of physical methods to solve chemical problems, and chemical physics,
in particular, developed in the 1930s as a true “borderland” discipline, as a response to
problems like the chemical bond, which had collaborative value for both chemists and
physicists.433
The discoveries Pauling made were not strictly physical in nature, yet could not
have been accomplished by purely chemical means, and for this reason it is the physical
foundation of the valence bond treatment that has been very provocative in the
philosophy of chemistry. As raised earlier in chapters 1 and 3, since Heitler and London
treated the homopolar bond in hydrogen as a purely physical system, and were largely
unaware of Lewis’ work, their research suggested that the roles of chemical combination
may be derived without the use of chemical methods.434 However, the discoveries
Pauling made were not strictly physical in nature, and his application of resonance
structures was accepted in the chemical community as an extension of Lewis’ previous
discoveries.
These two themes, interdisciplinarity and reductionism, form the core issues for
the chemical bond in history and philosophy of science, respectively, yet they create a
contrast. How do we resolve the fact that the same work of science that is historically
432 Servos, Physical Chemistry from Ostwald to Pauling, Gavroglu and Simões, “The Americans, the Germans, and the beginnings of quantum chemistry”; Nye, From Chemical Philosophy to Theoretical Chemistry; Park, Computations and Interpretations. 433 Park, Computations and Interpretations. 434 See Kohler, “The Lewis-Langmuir Theory of Valence and the Chemical Community”. Kohler argues that Kossel’s work more well-known in Germany than Lewis’, and so the theory of electron-pairing was not as well-known there as in the United States.
147
embedded in an interdisciplinary, physico-chemical “borderland”, has also been seen as
the source of an argument for theory reduction from chemistry to physics?435 Modern
philosophers of chemistry such as Andrea Woody, Eric Scerri, and Nikos Psarros have
argued an anti-reductionist standpoint, by emphasising the ways in which chemical
practice makes chemistry autonomous from physics, show that there is more to the
practice of chemistry than can be captured by the mathematics of the wave equation.436
From the historical perspective, traditional reductionism is quite limited because it
neglects the important developments made by chemists in chemical bond research, and
misrepresents the value of the valence bond treatment in chemistry. The issues of
interdisciplinarity surrounding the chemical bond suggest there is more to the story. The
valence bond had its origins in chemistry, but because of the physical treatment of Heitler
and London, we must also acknowledge its physical character. In this chapter I argue for
a reinterpretation of the chemical bond as a theoretical work that achieved a physico-
chemical synthesis, rather than a reduction of chemistry to physics.
This story of the chemical bond as a physico-chemical synthesis involves a
diverse set of protagonists within both the valence bond and molecular orbital methods,
in both of which wave mechanics is applied to the problem of bonding. After showing the
importance of shifting to interdisciplinarity from reductionism, this chapter will present
the work of two of Pauling’s American contemporaries, chemical physicists John Clarke
Slater and Robert Sanderson Mulliken, who have until now been limited to supporting
roles. Slater, who independently proved the tetrahedral structure of the carbon atom via
435 Park, Computations and Interpretations. 436 Woody, “Putting Quantum Mechanics to Work in Chemistry”; Scerri, “Has Chemistry at Least Been Approximately Reduced to Quantum Mechanics?”; Psarros, “The Lame and the Blind”.
148
the new methods in 1931, made many formal improvements to valence bond methods.437
His contributions led to valence bond theory becoming generally known as the Heitler-
London, Slater-Pauling (H-L-S-P) theory of valence. At around the same time that
Pauling and Slater made their first discoveries, Mulliken developed, with Friedrich Hund,
an alternative approach to chemical bonding called the molecular orbital method.
Through Pauling’s efforts the valence bond method became widely used until about the
1950s, when the molecular orbital method achieved wider acceptance following
prolonged disputes between supporters of the rival theories.438
Slater and Mulliken formed separate traditions of chemical bond research from
that of Lewis and Pauling, within the broader discipline of chemical physics. Their work
had much in common with Pauling’s, but create distinct paths of research because of their
different disciplinary training (Slater) and theoretical commitments (Hund-Mulliken) they
brought to their chemical bond studies. Later cultural divisions within chemical physics
created by disputes over the molecular orbital and valence bond methods intensified the
differences between the two theories, despite the fact that Mulliken found much to
identify with in the work of Lewis and Langmuir. Slater made his discovery of
hybridization within a physical tradition, and separated his work from Pauling’s. The
heterogeneous disciplinary nature of chemical physics, as a convergence of diverging
traditions, preserved these differences in thought in chemical bond research.439 Lastly,
the reception of Pauling’s work in the academic community, especially by
437 Park, “The contexts of simultaneous discovery”, 451-474. 438 Stephen G. Brush, “Dynamics of Theory Change in Chemistry. Part 1: The Benzene Problem, 1865-1945”, Studies in the History and Philosophy of Science, 30 (1999), 21-79; Karachalios, “On the Making of Quantum Chemistry in Germany”; Shaik and Hiberty, “Valence Bond Theory, Its History, Fundamentals, and Applications: A Primer”. 439 Gavroglu and Simões, “The Americans, the Germans, and the beginnings of quantum chemistry”.
149
contemporaries like Slater, confirmed the value of chemical bond research in an
emerging disciplinary climate. Concluding with a discussion of a series of job offers
Pauling received from 1929 to 1931, we see that by the early 1930s the American
academic climate had begun to shift in response to the “new” physical chemistry.
2. Reductionism vs. interdisciplinarity
Recently, philosophers of chemistry have studied the consequences of the
reductionist claim by challenging the degree to which chemical phenomena can be fully
explained predictively using quantum mechanics. Their arguments have questioned the
value of reductionism for understanding the relationship between physics and chemistry,
and its usefulness in view of the difficulties in solving the Schrödinger equation.440
Alternate approaches have suggested ways to understand the connection between physics
and chemistry without drawing on theory reduction, by focusing on the value of physics
and chemistry as consulting practices that may inform each other’s research questions,
and the possibility that chemists may adopt an autonomous perspective from physicists
regarding common entities of study.441 The first group raise concerns for the viability of
arguments of theory reduction for physics and chemistry, while the second suggest
alternate perspectives to reduction, presenting tentative steps in describing the
disciplinary significance of quantum chemistry.
Much of this debate has focused on the solvability of the Schrödinger wave
equation, which forms the basis of the valence bond treatment. Eric Scerri directly
440 Woody, “Putting Quantum Mechanics to Work in Chemistry.”, Scerri, “Has Chemistry at Least Been Approximately Reduced to Quantum Mechanics?”. 441 Psarros, “The Lame and the Blind”, Jenkins, Zach, “Do You Need to Believe in Orbitals to Use Them?: Realism and the Autonomy of Chemistry.” Philosophy of Science, 70 (2003), 1052-1062.
150
confronts the applicability of theoretical reductionism by examining the degree to which
approximation methods must be used in order to obtain a full solution. He suggests that in
the case of such problems as the chemical bond, which are not fully solvable by ab initio
means, it would be more appropriate to apply criteria of approximate reduction, in which
the relationships proposed by the reducing theory “hold not exactly, but with a certain
specifiable degree of error.”442 However, we cannot define accurately enough the degree
of error in the approximations because there is no independent ab initio criteria for
evaluating the calculations. Therefore, claims of complete theory reduction can only
mislead us based on the “apparent” successes of wave mechanics.
Woody is more optimistic about the viability of ab initio methods in principle
because the James-Coolidge calculation was successful in fully deducing the structure of
the hydrogen molecule: “[a] fact from one domain of inquiry was captured completely by
the theoretical structure of another domain.”443 The barrier she sees to accepting
reduction in practice is the modern reliance on semi-empirical methods to compensate for
the computational constraints methods like CI. These approximations form such a major
component of the calculations that they prevent purely theoretical methods from guiding
chemical practice. More important information for chemists, Woody argues, is that
conveyed through diagrammatic methods, such as molecular orbital diagrams, which can
be “reliably reproduced and recognized”.444 Such methods provide chemists with useful
ways of describing chemical structures that cannot be captured through purely quantum
mechanical methods, whether approximate or not.
442 Scerri, “Has Chemistry at Least Been Approximately Reduced to Quantum Mechanics?”, 168. 443 Woody, “Putting Quantum Mechanics to Work in Chemistry,” S615. 444 Woody, “Putting Quantum Mechanics to Work in Chemistry,” S624.
151
A common theme in modern anti-reductionist arguments is the autonomy of
chemistry from physics, seen in the different perspective that chemists bring to physical
entities and concepts, and in the ways chemical practice distinguishes itself from physics.
Zach Jenkins has studied the question of whether chemistry may be treated as
autonomous from physics concerning realistic interpretations of entities common to both
disciplines, such as atomic orbitals. He argues that modern methods such as CI do not
fully realistically describe atomic orbitals, because they neglect crucial parts of the whole
molecular system, such as the mutual interaction of electrons.445 As a result, chemists are
left with a picture that denies them information about real physical entities, such as “the
possible presence of bonds.”446 This creates the troubling scenario that chemists will
never be able to treat atomic orbitals realistically if they orbitals cannot be reduced to
entities described by the laws of physics. Scerri suggests that allowing chemistry an
autonomous viewpoint sidesteps such positions, implying that “chemistry can make use
of fundamental physics by borrowing such terms as orbits and configurations but would
not be required to follow strict physical laws to the point of denying the existence of
atomic orbitals in chemistry.”447 From these studies we see that the adoption of
autonomous perspectives from physicists can be an accepted practice in chemistry,
allowing for the possibility of autonomous viewpoints on other theories common to both
disciplines.
From Woody, we see that the discoveries of theoretical physics will never guide
chemical practice in the way they guide physicists, and Jenkins argument implies that
445 Jenkins, “Do You Need to Believe in Orbitals to Use Them?”, 1053-55. 446 Jenkins, “Do You Need to Believe in Orbitals to Use Them?”, 1053-56. 447 Scerri, “Philosophy of Chemistry”, 524,
152
chemical methods may always be physically limited.448 These arguments suggest there is
something about chemical practice that the laws of physics cannot capture, but more
importantly for the question of interdisciplinarity, that chemists must also rely on some
physics in order to achieve a complete physico-chemical understanding of the entities
they study. This is a key issue when we consider the climate of chemical physics in
which the chemical bond developed. As a discipline formed from the synthesis of
elements from previously distinct traditions, we must accept that physical and chemical
methods must be combined in order to solve the problems chemical physics entails: the
rejection of either the physical or chemical influence cannot help us understand the
chemical bond’s importance. However, moving too deeply down the path of chemical
autonomy would be equally dangerous. We must preserve the uniquely chemical aspects
of the chemical bond’s history, while acknowledging its physical components.
In the years following Heitler and London’s discovery, shades of reductionism
could even come from chemical contributors like Pauling. In remarks to a meeting of the
American Association for the Advancement of Science in 1928, he spoke of science as a
pursuit that “progresses more or less uniformly and monotonously toward a goal: the
reduction of the phenomena included within its domain to the simplest possible form”,
adding that theoretical physics had since the discovery of quantum mechanics “taken
mighty strides in this direction.”449 He felt, as many physicists did, that theoretical
chemistry now seemed to depend entirely on the Pauli exclusion principle, and the
phenomenon of resonance. As a young scientist who was using those tools to carve out
his early career as a chemical physicist, Pauling found the idea of reductionism exciting.
448 Woody, “Putting Quantum Mechanics to Work in Chemistry”. 449 Pauling, Linus, “NOTCB”, notes for AAAS meeting, June 14, 1928, p 1. Folder LP Speeches 1928s, LPP.
153
Park shows, however, that this attitude diminished later in Pauling’s career, in the 1930s,
as he learned the practical complexities of the Schrödinger equation and the value of
experimental work in interpreting solutions to valence bond problems.450 Mary Jo Nye
has suggested that this view simply reflected a physical “mode of thought” in Pauling’s
career, in contrast to the biological mode that occupied his later career, rather than
signifying a rallying cry for physical laws to replace chemical principles.451
When Pauling reflected on the matter later in his life, he concluded that the true
value of the resonance theory of chemistry had progressed so far beyond the precise
quantum mechanical calculations on which it was based, that it had really had the effect
of guiding empirical discoveries with simple principles.452 In this sense he believed
resonance was being used to continue a tradition in structural chemistry that had begun
long before physicists entered the picture. In this there is confirmation of Woody’s
argument that, despite any initial hope for the possibility that quantum mechanics
presented, pure physical theory is limited in the ways it can shape chemical practice. This
seems an apt belief for a chemist whose formative years were shaped by the foundational
work of quantum mechanics. Pauling was certainly happy about the impact theoretical
physics could have on chemistry. It is likely that his enthusiasm was motivated by an
initially naïve reductionism, in which he believed theoretical physics could guide the
future path of chemistry, through the solutions of problems like the valence bond. In the
context of his respect for Lewis’ program, such a physical mode of thought would be
expected within a physico-chemical synthesis.
450 Park, “Chemical Translators” 23-4. 451 Mary Jo Nye, “Physical and Biological Modes of Thought in the Chemistry of Linus Pauling”, Studies in the History and Philosophy of Modern Physics, 31 (2000), 475-491. 452 Nye, “Physical and Biological Modes of Thought”, 480.
154
In studying Pauling’s exposition of resonance for chemical audiences, Park has
argued that the successful appropriation of physical methods into chemistry was not
indicative of a reduction of chemistry to physics, but rather a sign that chemists had
translated physical knowledge into their own language, to make it useful in chemistry.453
Pauling and his collaborator, chemical physicist George Wheland, encountered
difficulties in teaching the theory of resonance, because it was often conflated with
tautomerism in the chemical community. Although Wheland believed resonance
structures were simply artificial, man-made mathematical constructions, Pauling believed
that the distinction between the two was more subtle, with resonance being another, more
quickly moving, form of tautomerism. As they debated the issue, Wheland and Pauling
gave the community a greater understanding of what resonance meant for chemistry, and
allowed chemists to “make use of quantum mechanics in their own language”.454
Therefore once resonance was translated in chemically meaningful terms, it became
separated from the original physically defined phenomenon, and allowed chemists to take
a step further into autonomous territory.
In confronting reductionism from a historical perspective, Ana Simoes and Kostas
Gavroglu suggest that it may be a “misplaced category” which expresses a trend in the
physicists’ culture that was not evident in the chemists’.455 Chemists were impelled to
respond to reductionist claims for fear of losing control of their own discipline, and
pitched mathematics as the chief factor that prevented reductionism from being realized:
since quantum chemistry was a highly mathematical pursuit, this was seen to buffer the
453 Park, Buhm Soon, “Chemical Translators”. 454 Park, Buhm Soon, “Chemical Translators”. 455 Simões and Gavroglu, “Issues in the History of Theoretical and Quantum Chemistry”, on 52.
155
claims of physical reduction.456 Other chemists (such as Pauling) diminished the
importance of mathematical derivations while teaching the concepts of chemical bond
theories,457 to emphasize their chemical utility. Still other quantum chemists played up
the mathematical nature of their own research in order to distinguish their discipline from
theoretical physics.458 In their research, Gavroglu and Simoes highlight how the prospect
of reductionism inspired chemists to work quickly to integrate the new quantum
mechanics into their research, while still making it their own, lest it be mistaken for
physics.
Recent historical studies of the development of the chemical bond during the
1920s and 1930s have raised greater awareness of the disciplinary character of the
research into chemical bond theory, quantum mechanics and quantum chemistry.
Chemical physics emerged as a result of problems that had raised questions in many areas
of the physical sciences, in contrast to physical chemistry, which by the early twentieth
century became largely a sub-field of chemistry carried out by chemists only.459
Gavroglu and Simoes have emphasized the heterogeneous makeup of quantum chemistry
within chemical physics because of the different disciplinary origins and national styles
of its contributing scientists.460 Buhm Soon Park typifies quantum chemistry as the
collected work of individual contributions, with a diversity that was celebrated by the
members of its field. These scientists, unlike the older physical chemists, could claim
456 Gavroglu, and Simões, “Preparing the ground for quantum chemistry in Great Britain”. 457 Simões and Gavroglu, “Issues in the History of Theoretical and Quantum Chemistry”. 458 Gavroglu and Simões, “Preparing the ground for quantum chemistry in Great Britain”. 459 Servos, Physical Chemistry from Ostwald to Pauling; Nye, From Chemical Philosophy to Theoretical Chemistry. 460 Gavroglu and Simões, “The Americans, the Germans, and the beginnings of quantum chemistry”.
156
their work was well situated with the “borderland between physics and chemistry.”461 It
was in this transitional period that the chemical bond developed, by contributing to the
formation of chemical physics as a physico-chemical problem studied by physicists and
chemists.
A drawback of philosophical arguments is that although Woody and Scerri show
the limited potential of reductionism, their analyses tend to focus on the modern status of
quantum chemistry, which leaves us with very inadequate means of analysing the
repercussions of reductionism for the chemical bond as a historical entity. Another
weakness of current accounts of theory reduction, as Woody has argued is that they
provide few useful ways of understanding the connection between the ‘reduced’ (here,
chemistry) and ‘reducing’ (here, physics) theories, because in most accounts the
‘reduced’ theory disappears.462 The valence bond provides us with an opportunity to
explore the connection between physics and chemistry, but we must do more than offer
the perspective that chemists and physicists view theoretical entities in different ways: in
the climate of a physico-chemical “borderland”, it is impossible for any work of
chemistry to be autonomous of physics. If we are to appreciate the full significance of
the Heitler-London treatment we must understand its relevance as a work of chemistry, as
well as a work of physics. Chemical autonomy presents the surest barrier to reductionism,
and is exemplified by both the philosophical and historical threads of the counter-
arguments. We must incorporate this into our understanding of the interdisciplinary
context of the chemical bond’s origins.
461 Park, Computations and Interpretations, iv. 462 Woody, “Putting Quantum Mechanics to Work in Chemistry”, S625-26,
157
The example of Lewis and Pauling provides evidence of a tradition of chemical
bond research that supported an autonomous perspective for chemists regarding the atom
and the chemical bond, within a synthesis of physical and chemical theory. Lewis’
postulation of a chemical atomic model was directed toward the chemists’ needs for a
theory of bonding. The model’s acceptance within the chemical community, despite the
widespread physical use of the Bohr atom, reflected an autonomous chemical viewpoint,
which Pauling preserved in The Nature of the Chemical Bond. Here, chemical autonomy
allows us to acknowledge the physic-chemical nature of the chemical bond, by
addressing its physical basis without negating its chemical origins. Instead of examining
the interdisciplinary context of the chemical bond, in terms of the nature of chemical
physics, addressing the valence bond through Lewis and Pauling allows us to study the
interdisciplinary character of the science itself.
Reductionism is undoubtedly a challenge for philosophers and historians of
chemistry, but within the philosophical and historical perspectives is a tension between
the reductionist claim and the inherent issues of interdisciplinarity that have made the
chemical bond, physical chemistry and chemical physics so historically fascinating.
Given the importance of chemical autonomy in anti-reductionism arguments, a more
profitable response to reductionism is to acknowledge the respective claims of chemistry
and physics to the chemical bond. The philosophical and historical perspectives lead us to
a contrast: the valence bond treatment has been seen as the basis for the reductionist
argument, while historians have seen it as a gateway to disciplinary convergence. While
so much of the philosophical debate surrounding the chemical bond has focused on
reduction, separating this issue from its historical importance as an interdisciplinary
158
entity prevents us from understanding the chemical bond’s full significance. By situating
the valence bond in its chemical context, we see that it is best described as a physico-
chemical entity that is best represented as a theoretical synthesis.
The following sections will address the broader context for this synthesis, in two
ways: first, by presenting two alternate cases of chemical bond research in the same
period as Pauling’s The Nature of the Chemical Bond study, and second, his reception in
the emerging chemical physics community. We have already seen glimpses of the
institutional interest in Pauling’s new interdisciplinary directions, with his acceptance in
the Berkeley chemistry community. The surrounding contributions to chemical bond
theory from scientists outside Pauling’s immediate circle, within both physics and
chemistry, illustrate the diversity of responses to the chemical bond in the disciplinary
context. As the complexity of the chemical bond as a theory provides a barrier to theory
reduction, the complexity of responses in the community likewise show that the chemical
bond is a problem that both requires the support of, and attracts, a physico-chemical
disciplinary framework.
3. Alternate avenues: the chemical bond elsewhere
3.1 John Clarke Slater and the valence bond
John Clarke Slater was born in 1900, and spent the years of his formative science
education in the late 1910s and early 1920s, as part of what he called the “lucky
generation” of physicists. 463 Too young to have had World War I seriously interrupt their
academic life, and emerging fresh from the quantum theory to start their careers just as
463 Schweber, “The Young John Clarke Slater”, Historical Studies in the Physical Sciences, 20 (1990): 339-406, on 339.
159
relativity and quantum mechanics entered the scene, they had a sense of limitless
possibility of what American science could become. After completing an undergraduate
degree in physics at the University of Rochester, going to Harvard for graduate work
“seemed like the obvious thing to do” to Slater, with Harvard being “one of the places
that really were doing modern physics.”464 As was traditional for many American post-
graduates, he spent a year in Europe after completing his graduate work in 1923. Slater
had expected to spend this time with Bohr, but found it to be something of a let-down:
instead of arriving to a wealth of new ideas, he found he had already learned as much
about quantum theory in the United States as the Europeans had.465 Disillusioned with his
experiences abroad, Slater returned to his home country with hopes of building up
physics there, so those Americans who followed the “lucky” generation would not need
to depend on Europe to complete their training.
In the 1920s and early 1930s, Slater held faculty positions at first Harvard and
then M.I.T., and worked at both institutions on raising the status of theoretical physics in
the United States. At Harvard he made plans to renovate and update the Jefferson
Laboratory of Physics to integrate laboratory and theoretical research with teaching. In
the same way Pauling had advocated bringing wave mechanics into programs of physical
chemistry, Slater did the same for his physics departments, and worked to develop
quantum chemistry and theoretical physics in collaboration with departments of
chemistry. At the time Slater moved from Harvard., M.I.T. President Karl Compton
become suitably impressed by the strides Noyes group had made at Caltech and was
keenly aware of M.I.T.’s comparative shortcomings. When M.I.T. had fallen behind after
464 Schweber, “The Young John Clarke Slater”, 348. 465 Schweber, “The Young John Clarke Slater”, 351.
160
Noyes’ departure, Caltech had become the “premier” American institution for “training
scientists and engineers.”466 Slater’s leadership of physics at M.I.T. established a
curriculum that showed applying quantum mechanics to the properties of matter was the
new direction for physics. In Schweber’s view, Slater’s work was so successful in
establishing quantum chemistry as an academic discipline in American universities that it
helped make it into a “quintessentially American” field.467
Though categorized historically as a quantum chemist, Slater was a physicist by
training and in practice, and in his early career bridged the era of quantum theory with the
discovery of quantum mechanics. Like Pauling, who had first studied the chemical bond
in the context of the quantum theory before learning wave mechanics in his postdoctoral
year, Slater had also first approached the quantum properties of matter in the context of
the older quantum theory. Still in the formative years of his career when wave mechanics
broke on the scene, Slater saw it as a simple extension of the quantum theory he had
already become expert in. In 1925, independently of Uhlenbeck and Goudschmidt, he
advanced the idea of a fourth quantum number, and latched on quickly to many-body
problems.468 Slater’s major scientific contributions from this period came between 1929
to 1931, during which time he developed the determinantal method (see following
paragraph) and discovered orbital hybridization in methane independently of Pauling.
With this work Slater was interested in improving on Heisenberg’s work and applying the
formalism of wave mechanics to better understand the physical state of matter: the
molecular state and other problems in chemistry presented him with many opportunities
466 Schweber, “The Young John Clarke Slater”, 363. 467 Schweber, “The Young John Clarke Slater”, 341. 468 Schweber, “The Young John Clarke Slater”, 371.
161
for research.469 In the same year Pauling made his first published forays into the valence
bond problem, Slater was applying the formalism of wave mechanics to problems like the
helium atom.470
Slater’s determinantal method was a generalization of the Heitler-London
treatment to a system of n electrons, which expressed the total wave function as a product
of bond wave functions of the type Heitler and London had used.471 Many-electron
systems had become more complex with the extra constraint of spin. Heisenberg had
found that for anti-symmetric wave functions, it was impossible for two or more electrons
to exist in the same orbit, which was essentially what the exclusion principle stated.472
This allowed for the interpretation that two antisymmetric wave functions, with spin up
and down, respectively, formed the bond. Many European physicists, led by Heitler,
tackled this with group theory, by abstracting the problem to the possible permutations of
spin arrangements that could make up a molecule.
Slater was dissatisfied with group theory, and found it unintuitive. Instead, he
proposed a solution that treated the bonding electrons as functions of their coordinates (n,
l, ml, ms), which included spin as the fourth number. The product of all these functions
gave a symmetric, approximate solution to the Schrödinger equation. Slater’s innovation
was to apply matrix arithmetic and construct a determinant out of the individual electron
functions, which would produce the desired antisymmetric wave function. This then
satisfied the exclusion principle and gave an approximate solution to the wave equation.
By 1930, although he had not yet made a rigorous calculation of the bond energy, and
469 Park, “The Contexts of Simultaneous Discovery”, 453. 470 John C. Slater, “The Normal State of Helium”, Physical Review, 32 (1928), 349-360. 471 Shaik and Hiberty, “Valence Bond Theory”, 4. 472 Park, “The Contexts of Simultaneous Discovery”, 453; Schweber, “The Young John Clarke Slater”, 453.
162
could not treat directed bonds, Slater was now confident in his ability to approach
molecules with multiple bonds using the determinantal method.473 Pauling was one of the
first to sing the method’s praises, particularly of Slater’s incorporation of spin functions
which had previously been difficult to visualize.474 Many physicists, “especially those
who saw group theory as an arcane, incomprehensible mathematical manipulation,”475
welcomed the theory with open arms.
As a work of science, the determinantal method was clearly theoretical physics:
Slater attacked the problem of bonding from a formal perspective, to find a more elegant
way of approaching the problem of a chemically bonded structure. His improvements
pushed the study of chemical bonding a great step forward by removing layers of
complexity from a problem that had been challenging from the start. Pauling began to
incorporate Slater’s discoveries into his lectures, as he produced his own formative work
on the valence bond. “The most striking and pleasing characteristic” of Slater’s methods,
Pauling told his audience at Berkeley in 1932,
[is] its simplicity and lucidity. With the aid of only elementary and straight-forward arguments he has not only obtained results equivalent to those got by … methods (mainly involving group theory) that were so involved and abstract that I, for one, despite much study was never able to say that I had got a feel for them, but has also gone far beyond the point they reached, introducing a new era in the application of quantum mechanics to molecular structure. I might mention that everyone is hoping that before too long he will introduce still another era by discovering a treatment which will permit the actual numerical calculations to be made without the very considerable difficulties that attend them now.476
473 Park, “The Contexts of Simultaneous Discovery”, 453-8. 474 Linus Pauling, Fifth Berkeley Lecture, “Slater’s Method of Formulating Wave Functions for Molecules”, May 12, 1932, 19. LP Speeches 1932a, LPP. 475 Park, “The Contexts of Simultaneous Discovery”, 454. 476 Linus Pauling, Fifth Berkeley Lecture, “Slater’s Method”, May 12, 1932, 17
163
Pauling’s praise for Slater’s work showed the great respect he had for his contemporary,
and also the suggested that Slater had a natural intuition and facility for the theoretical
treatment of the chemical bond as a physical system. Rather than viewing the bond as a
reduced entity, Slater’s expertise with the physical formalism of the chemical bond aided
chemists like Pauling, as well as physicists, who were researching bonded structures.
Slater’s study of orbital hybridization embodied much the same character as his
determinantal method.477 Initially, he approached methane as another application of the
determinantal method, as a sequel to a 1930 study of metals.478 Like Pauling, Slater
visualized the methane molecule in the shape of a pyramid, with the four single bonds to
hydrogen atoms formed through the linear combination of one s and three p wave
functions.479 In this first paper, a brief 9 pages, he set out his proposed method, and
presented different examples of divalent, trivalent, and tetravalent atoms, the last of
which “demand a different treatment” from the rest.480 It was in his following paper481
that he gave the full theoretical justification for the work, showing how wave functions
would be set up from the determinantal method and discussing what kind of integrals
would be computed.
This theoretical paper laid out a formal procedure for “solving the Schrödinger
equation for molecules in general, rather than to carry out computations for actual
examples.”482 With a comprehensive treatment of the valence bond as a physical system,
Slater began with the details of his method and working with the individual equations,
477 John C. Slater, “Directed Valence in Polyatomic Molecules”, Physical Review, 37 (1931), 481-489. 478 John C. Slater, “Cohesion in Monovalent Metals”, Physical Review, 35 (1930), 509-529. 479 Slater, “Directed Valence in Polyatomic Molecules”, 486. 480 Slater, “Directed Valence in Polyatomic Molecules”, 485 481 John C. Slater, “Molecular Energy Levels and Valence Bonds”, Physical Review, 38 (1931), 1109-1144. 482 Park, “Contexts of Simultaneous Discovery”, 461.
164
and moved on to examples and special cases: for example, systems with two atoms and
one s electron each, or bonding with three electrons to create spin degeneracy. Methane,
the case that made Slater and Pauling so well-known, was presented here as one example
of a system where five atoms were involved in bonding, with four s and one p shell
involved in forming the molecule.483 The formal treatment gave the directional properties
of valence, which had not been obtained in the previous physical approaches, and it is
because of this important step that the discovery of hybridization has been regarded as the
beginning of quantum chemistry.484
As Park and Schweber have emphasized, Slater introduced the concept of directed
valence nearly simultaneously with Pauling in 1931, through entirely independent lines of
research in the previous years.485 Pauling had first made mention of it in his short paper
for the National Academy of Sciences in 1928, while Slater had presented it to the
American Physical Society in April of 1930. No priority dispute arose between the two,
and they were fully respectful of each other’s contributions to the same problem. In his
first Berkeley lecture of 1931, Pauling presented his own results on hybridization, such as
the explanation of tetrahedral valency and lack of free rotation about double bonds, which
in turn were about to be published in the Journal of the American Chemical Society the
following month. In this lecture he noted very plainly that some of the results “were
independently obtained by Slater, who has published a preliminary note in the March 1
Physical Review.”486
483 Slater, “Molecular Energy Levels and Valence Bonds”, 1140. 484 Karachalios, “On the Making of Quantum Chemistry in Germany”, 494. 485 Linus Pauling, “The Shared Electron Bond”, Proceedings of the National Academy of Sciences, 14 (1928), 359-362; Schweber, “The Young John Clarke Slater”, 387; Park, “The Contexts of Simultaneous Discovery”, 458-9. 486 Pauling, “The Nature of the Chemical Bond”, First Berkeley Lecture, March 23, 1931, 2. LP Speeches 1931a, LPP.
165
Slater was equally complimentary about Pauling whenever their work approached
similar territory, and Pauling’s records show only friendly relations between them. A
letter from Slater to Pauling in August of 1930 suggests that, while the two men had not
collaborated on any of the work, they were keeping abreast of one another’s work,
reflecting the interest Slater had shown in Pauling’s career since he made him a Harvard
job offer in 1929 (see section 4). Upon completing his 1930 paper on metals, Slater wrote
to his colleague, “I think you will find that it supplements your [work] nicely, without
overlapping at all.”487 The following year, when they had both published on directed
valence, Slater wrote to Pauling in anticipation of a meeting in Pasadena which they were
both attending, to pass on his compliments. “I am glad things worked out as they did”,
Slater wrote, “we both deciding simultaneously to write up our ideas.… Incidentally, our
general points of view seem so similar that we shall want to compare notes before the
meeting, to avoid saying the same things.”488 On speaking of Pauling to a colleague,
Slater remarked on how much they had in common: “although he calls himself a chemist,
the things he actually does are almost exactly in the same line I am working in, and it
seems we could cooperate very well.”489
Further exemplifying the contrasting perspective Slater brought to the valence
bond was the way he presented some of the concepts in his published work. His style was
more formal and at times more precise than Pauling’s, since his papers were pitched
toward the audience of the Physical Review rather than the chemical audience who were
still learning the intricacies of quantum mechanics. In particular, owing to the nature of
the determinantal method, Slater’s definition of the homopolar bond put more emphasis
487 Letter from Slater to Pauling, August 8, 1930, LP Safe 2.005, LPP. 488 Letter from Slater to Pauling, April 4, 1931, LP Safe 2.005, LPP 489 Slater, quoted in Schweber, “The Young John Clarke Slater”, 481.
166
on the spin function and how it could explain the equilibrium of attraction that occurs
when two electrons of opposite spin overlap.490 Like Pauling, though, (and Lewis,
besides), Slater believed that the two classes of polar and non-polar were “not as a matter
of fact very different in fundamentals”, and the polar transfer of electrons could be seen
as an extreme case of sharing as in the homopolar type. Further, the shared-pair method
could “deal in a general way with directional properties.” 491 To Slater the “phenomenon
of shared valence” was “of common occurrence”, and he felt that “an understanding of it,
as well as the relation between ionic and homopolar binding, is essential.”492 Although
there were disciplinary differences in the way they approached the problem of valence
bonding, Slater and Pauling employed their different approaches to arrive at similar
conclusions
Both men may be categorized as chemical physicists, but Slater, trained as a
physicist, and Pauling, trained as a chemist, brought different backgrounds to their
treatments of the directed valence problem. Slater’s goal was to lay out a theoretical
procedure for the solution of molecules in general. In general, his work was directed
toward a physical audience who desired formal methods of understanding the structure of
matter. Pauling, by contrast, derived many of his conclusions from his experimental
understanding of bond energies, atomic radii, and bond angles, obtained from his studies
with x-ray crystallography. Knowing his chemical audience would not welcome lengthy
derivations like those Slater gave for his Physical Review readership, Pauling reduced his
own proofs to problems like directed valence to general sketches.493 Using visual
490 Slater, “Directed Valence in Polyatomic Molecules”, 481. 491 Slater, “Directed Valence in Polyatomic Molecules”, 488. 492 Slater, “Directed Valence in Polyatomic Molecules”, 489. 493 Park, “Contexts of Simultaneous Discovery”, 468-70.
167
interpretations of resonance and interpreting rules from the existing canon of quantum
mechanics, Pauling presented an appealing system of structural chemistry, which he was
“confident that chemists would like”.494
Despite the similarities in Slater and Pauling’s work, and their independent
solutions to identical problems of chemical structure, their contributions to valence bond
theory were formed in distinct traditions of research. Park suggests that although Slater
and Pauling actively promoted interdisciplinary research, their respective physical and
chemical disciplinary ties allowed them both to claim simultaneous discovery without
dispute over priority.495 The only disagreement between them was over method: Park
suggests that Slater accepted Pauling’s work as an independent verification of the same
phenomenon he had discovered, but did not agree with his quantitative arguments.496 The
fact that these two men reached a “common destination” from their separate paths is
indicative, Park argues, of the tensions between physics and chemistry during a time
when the two disciplines were beginning to experience a substantial overlap.
As shown in the work of Lewis and Pauling in bringing together quantum physics,
atomic theory, and chemical valency, interdisciplinarity is inherent in chemical bond
research. However, distinct traditions and allegiances may still be carved out within this
body of work. Rather than becoming a melting pot of convergence, the interdisciplinarity
of chemical physics in the 1930s was a synthesis in which the physical and chemical
identity of its contributors was exemplified by the way in which they carried out their
work. Though Slater and Pauling made equal contributions to a foundational problem in
chemical physics, a discipline which they helped to found, their scientific upbringing as
494 Park, “Contexts of Simultaneous Discovery”, 470. 495 Park, Computations and Interpretations. 496 Park, “Contexts of Simultaneous Discovery”
168
physicist and chemist, respectively, and their preferences for methodology that appealed
to those audiences, remained distinct in this interdisciplinary climate.
3.2. Molecular orbitals and Robert Mulliken
A second example of an alternate tradition of chemical bond research within
quantum chemistry can be found in the theory of molecular orbitals, developed by
German Friedrich Hund and American Robert Mulliken. Many authors have previously
addressed the rivalry between the molecular orbital and valence bond theories from the
1930s to 1960s, and the cultural divides the debate created within the chemical physics
community.497 Both theories employed wave mechanics to model the behaviour of
bonding electrons, but the valence bond approach was easier to visualize, and was a
better intuitive fit with the way most chemists thought of molecules, as a collection of
linked individual atoms.
The dichotomy of molecular orbital and valence bond theory presents a case of a
cultural fracture within chemical physics created by opposing approaches to the same
problem. Molecular orbital theory, however, shared many conceptual ties with the
foundations of valence bond theory, and with the work of Lewis and Langmuir, and
exhibited the similar trait of being an interdisciplinary theory formed from the work of a
chemist and a physicist. Despite these parallels, and the fact that the two theories of
chemical bonding are now regarded as complementary approaches to a singular problem,
they have been long held historically as different traditions of chemical bond research.
What influenced chemists’ allegiance to a particular theory of bonding was their 497 Karachalios, “On the Making of Quantum Chemistry in Germany”, Simões and Gavroglu, “Issues in the History of Theoretical and Quantum Chemistry”; Shaik and Hiberty, “Valence Bond Theory, Its History, Fundamentals, and Applications”; Brush, Stephen G. “Dynamics of Theory Change in Chemistry”.
169
methodological and personal preferences, dividing the culture of the discipline along two
different physico-chemical lines.
Mulliken was born in 1896, and grew up in Newburyport, Massachussetts. He
came by an interest in chemistry at a young age, through his father, who was a high
school friend of Arthur Noyes. Like Noyes, Mulliken’s father went on to M.I.T., and then
Leipzig, for degrees in chemistry, and became a professor at M.I.T. When Noyes left for
Pasadena to direct the Chemistry Division at Caltech, he unsuccessfully tried to convince
Mulliken’s father to join him there as a professor. As Mulliken recalled in his memoirs,
“[m]y own life would have been different if we had moved.”498 Interested in both
chemistry and physics, and philosophy as well, Mulliken eventually decided to pursue
chemistry upon entering M.I.T. as an undergraduate.499 His earliest research there was an
assignment from Noyes on complex ions in solution, but Mulliken found it unsatisfying
and he turned quickly toward atomic structure. In his high school days he had latched on
to the problem of the electron, but by the time he came to do his PhD, having been
influenced by Rutherford and Bohr’s discoveries about the nuclear atom, it seemed to
him that “the subject that most deserved attention was that of atomic nuclei.”500 At that
time in the United States, the physical chemist William D. Harkins was the only one
studying this area, and so Mulliken went to work with Harkins at the University of
Chicago to complete his PhD in 1921. From there, he held a National Research Council 498 Robert S. Mulliken, Life of a Scientist: An Autobiographical Account of the Development of Molecular Orbital Theory with an Introductory Memoir by Friedrich Hund (Berlin: Springer-Verlag, 1989) (Bernard J. Ransil, editor), 8. Given the influence Noyes had on Pauling’s early development of the chemical bond, one can imagine how different the field of quantum chemistry might have been as well. 499 His choice to go to M.I.T. was made for financial reasons, as his support, from the Wheelwright Fund, favoured it. Mulliken said later he would have preferred to attend Harvard “with the hope of gaining broader viewpoints and even going into something different than science.” Mulliken, Life of a Scientist, 17. 500 Mulliken, Life of a Scientist, 26.
170
fellowship until 1926, and was Assistant Professor of Physics at Washington Square
College (New York University), before settling into a position in physics at the
University of Chicago in 1928.
Over the course of his time as a student and post-doctoral graduate, Mulliken met
with and studied the work of many of the men who would come to shape physico-
chemical work in the following decades. His first introduction to the quantum theory
came from Robert Millikan, who was “full of enthusiasm” for it, even though it seemed
to him to be “an awful mess”.501 From 1924 to 1926, in a house near Harvard Square, he
shared neighbouring rooms and much scientific discussion with John Slater, who had
then just returned from his year in Copenhagen with Bohr. Also at Harvard was John van
Vleck, who had recently completed his Ph.D. there, and J. Robert Oppenheimer, who had
received his undergraduate degree there, years before his later association with the
Paulings in Pasadena.
As an undergraduate Mulliken also became interested in Lewis’ work on the
chemical bond, and read Langmuir’s subsequent papers. At the time, though people
spoke of the “Lewis-Langmuir theory”, Mulliken concluded “it was mostly Lewis’
although Langmuir was rather assertive.” This first discovery of the chemical bond had a
memorable place in his biographical recollections. As Mulliken described it, the
experience of discovering this theory of the chemical bond gave him the feeling that
“with G.N. Lewis as a guide,” he could “learn more about what electrons were doing in
molecules, a subject which later led [him] to work which won a Nobel Prize.”502 He
remained inspired about the properties of electrons in atoms during his stay at Harvard,
501 Mulliken, Life of a Scientist, 32. 502 Mulliken, Life of a Scientist, 32.
171
and from there he went (much like Pauling and Slater) to Germany to research them
further, in the spring of 1927.503 While in Göttingen, he made the acquaintance of Hund
(who was lecturing on molecular structure there that summer) and began a series of
meetings and correspondence that formed the foundations of molecular orbital theory.
Despite Mullikan and Hund’s close association, the two men never wrote a paper
together, but instead “interacted and stimulated one another by visits and lively
discussions.”504 At Göttingen and after, they independently systematized their knowledge
of molecular spectra by assigning quantum numbers to individual electrons. Over the
following years, via letters and discussion when Hund visited Chicago in 1929, he and
Mulliken worked out nomenclature and a scheme of chemical bonding that assigned
quantum numbers to the electrons in molecules based on an experimental understanding
of band spectra and its interpretation in the new wave mechanics.505 From 1928 to 1932,
the same years in which Pauling developed his first work on the valence bond, Mulliken
prepared a series of papers in which he outlined the value of molecular orbitals in the
broader context of valence theory.
In Mulliken’s view, Lewis and Langmuir had laid a clear foundation for chemical
bond studies, but it was not necessary to build this work upon the notion of the electron
pair. As he discussed the major features of his theory and the previous and current
traditions of valence bond theory, Mulliken presented the Lewis-Langmuir version of
bonding as a set of rules that could now be reformulated in the language of wave 503 Mullikan, Life of a Scientist (second preface, by Friedrich Hund), vii. 504 Mullikan, Life of a Scientist, v. 505 For Mulliken’s presentation, which includes substantial references to contemporary papers on aspects of both molecular orbital and valence bond theory, see Robert S. Mulliken, “The Assignment of Quantum Numbers for Electrons in Molecules. I.”, Physical Review, 32 (1928), 186-222.; Robert S. Mulliken, “The Assignment of Quantum Numbers for Electrons in Molecules. II. Correlation of Molecular and Atomic Electron States”, Physical Review, 32 (1928), 761-772.; Robert S. Mulliken, “The Assignment of Quantum Numbers for Electrons in Molecules. III. Diatomic Hydrides”, Physical Review, 33 (1929), 730-747.,
172
mechanics, but by abandoning Lewis’ idea that a molecule consisted of “specific atomic
or ionic units held together by discrete numbers of bonding electrons.” 506, 507 Along with
the concept of electron pairing, Lewis had earlier introduced the notion that each nucleus
in the molecule was surrounded by electrons, and that shared electrons were localized
between the two atoms they connected. While the pair had a fundamental importance for
Lewis, and it had been a major point of debate for him and Langmuir in naming the
theory, to Mulliken the pairing was only an “incidental” rather than “really essential
characteristic of chemical combination.”508 Mulliken’s homage to Lewis’ work
demonstrates that on a conceptual level, molecular orbital theory shared foundational ties
to valence bond theory, despite its distinct approach to the same problem.
Perhaps reflecting Langmuir’s differences from Lewis, Mulliken showed much
more personal affinity for Langmuir’s version of the octet theory than Lewis’ original
statement, in his reviews of chemical theories of valence.509 In reformulating Lewis’
work, Langmuir had also abstracted the concept of the electron pair from the group of
eight, which Mulliken identified with. The octet rule had separated electronic
arrangement from the concept of fixed pairings, and Mulliken latched on to the way
Langmuir’s assigned electronic structures based on whole groups of eight, rather than
pair-by-pair. Isosteres were particularly interesting to Mulliken, as well as Langmuir’s
treatment of special cases on the basis of total valence electrons rather than the number of
bonding pairs. By the same token, Mulliken felt that Pauling and Slater had in turn over-
506 See Mulliken, “Electronic Structures of Polyatomic Molecules and Valence”, Robert S. Mulliken, “Electronic Structures of Polyatomic Molecules and Valence. II. General Considerations”, Physical Review, 41 (1932), 49-71. 507 Mullikan, “Electronic Structures of Polyatomic Molecules and Valence”, 57. 508 Mullikan, “Electronic Structures of Polyatomic Molecules and Valence. II.”, 57. 509 Mullikan, “Electronic Structures of Polyatomic Molecules and Valence. II.”, Robert S. Mulliken, “The Assignment of Quantum Numbers for Electrons in Molecules. I.”.
173
emphasised only one aspect of Lewis’ theory, that of shared pairing, because Heitler and
London had given them a method that allowed them to do so.510 This had led valence
bond practitioners to promote what Mulliken saw as a conceptually and computationally
weaker method, because it forced all molecules to be viewed, rather artificially, as a
conglomeration of atoms and ions.
In studying Mulliken’s early papers, it is not hard to see why the molecular orbital
theory was slower to gather support than valence bonding. His writings were methodical
and comprehensive, showing in thoughtful detail what he saw as the advantages and
disadvantages of contemporary chemical bonding theories. However, these qualities also
made his papers seem verbose and unstructured. When compared with Pauling’s clear,
easy style that made wave mechanics friendly for chemical use, Mulliken would have
seemed intimidating and inaccessible to American chemical audiences.
In the following decades, the two methods achieved varied success with various
types of structures, depending on which approach was used. Slater himself wrote to the
Physical Review in 1932 to clarify misconceptions about the differences between them
and to emphasise that molecular orbitals and valence bonds were “complementary, not
antagonistic” methods, which in the end simply reduced to “the same integrals”, in
practice.511 He personally felt that molecular orbital techniques offered distinct
advantages, that might work in its favour in the long term. In 1931, Erich Hückel found
that molecular orbital methods gave much more accurate approximation of the bond
energies in benzene, but even still Slater and Pauling’s treatments of methane were
510 Mullikan, “Electronic Structures of Polyatomic Molecules and Valence. II.”, 54. 511 John C. Slater, “Note on Molecular Structure (Letter to the Editor)”, Physical Review, 41 (1932), 255-7.
174
widely viewed as a success of valence bonding.512 It was in the 1950s that molecular
orbital methods became broadly preferred, particular for organic chemists who worked
with large aromatic molecules, for which the valence bond calculations involved a
laborious amount of resonance structures.513 Overall, viewing a molecule as a whole
entity created from the electrons of bonding atoms, instead of a joining of individual
atoms which had overlapped their electron shells, led to a method that gave better
agreement with experimental results and had better predictive power. By the 1960s,
molecular orbital theory became the widespread favourite.
Valence bond theory also began to accrue a series of “failures”, such as the
inability to predict paramagnetism in the case of O2.514 These cases were a rallying point
for molecular orbital supporters. By the 1970s, molecular orbitals began to supplant
rather than supplement valence bonding. Brush argues that molecular orbitals eventually
won because chemists found it easier to use, computationally, and it simply produced
better results.515 Shaik and Hiberty have recently argued that valence bond methods have
seen a minor renaissance in teaching chemical bond theory, however, because of its
superior visual accessibility with diagrams and models.
Mulliken’s own view of molecular orbital theory in the context of Lewis,
Langmuir, Pauling and Slater, provides an example of a distinct tradition of chemical
bond research within chemical physics that also owed a conceptual debt to Lewis. While
the molecular orbital and valence bond theories have been historically viewed as
competing methods which provoked cultural divisions in the chemical physics
512 Brush, “Dynamics of Theory Change in Chemistry: Part 2,” 264. 513 Brush, “Dynamics of Theory Change in Chemistry: Part 2”, 270; Shaik and Hiberty, “Valence Bond Theory, Its History, Fundamentals, and Applications”, 8. 514 Shaik and Hiberty, “Valence Bond Theory, Its History, Fundamentals, and Applications”, 9. 515 Brush, “Dynamics of Theory Change in Chemistry: Part 2”, 285-288.
175
community, they nevertheless share common ties to the same chemical ancestors. While
Pauling saw himself as applying physical methods to carry out Lewis’ program, Mulliken
believed there was more to be carried out than only the study of electron pairing. The
Nature of the Chemical Bond was not the only valuable outcome of Lewis’ work. Just as
Langmuir had identified with the octet rule more than the rule of two, Mulliken had
identified with his treatment of the whole molecule rather than the postulate of localized
bonding. But, the complexity of issues in chemical bond theories meant that Mulliken
could identify with Langmuir without identifying with the valence bond. His example
shows that chemical physics could follow many conceptual pathways when using
physical methods to chemical ends.
4. Chemical bond in demand: Pauling as job candidate
Accompanying these developments in chemical bond theory in the 1920s and
1930s was a gradual change in the disciplinary framework that supported physico-
chemical research in American universities. As academic departments became aware of
the impact of the new trends in research, particularly those related to quantum mechanics,
they introduced changes in curriculum and faculty in order to keep pace. The examples of
Pauling, Slater and Mulliken have illustrated the diversity of academic traditions that
could produce chemical bond research, allowing for different styles and differences in
theoretical approach. This final section will study in detail Pauling’s experiences in this
period of disciplinary growth, through a series of job offers he received from institutions
hoping to develop their physics and chemistry programs along his interdisciplinary lines.
Although Pauling accepted none of these offers, his response and his interaction with
176
other members of the community provides further evidence of the fragmented nature of
chemical physics during this time.
The first of these offers came from Harvard, just as Pauling’s first series of
lectures at Berkeley was underway, late in February of 1929. Harvard was looking to
hire a new associate professor of chemistry following the passing of T.W. Richards,
Lewis’ former supervisor and the first American Nobel laureate in chemistry.516 While
Pauling had not originally been at the top of their list, Harvard was interested in his
experimental and theoretical research.517 The Chemistry Division there had recently
completed the construction of a new laboratory, and Gregory Baxter, the head of the
Division, believed it would be suitable for Pauling’s work. The hope was that Pauling’s
primary teaching would be an advanced course in physical chemistry, so he could bring
his talents in the ‘new’ wave-mechanically influenced physical chemistry to the Harvard
community.518
Harvard’s offer was a serious one, and Noyes experienced a fair amount of grief
over the possibility of losing Pauling. One of his organic chemists, James B. Conant, had
recently accepted a position there, having been offered inducements “quite without
precedent”, including “freedom from teaching and administration and liberal financing of
his researches” in addition to the fine new laboratory Noyes worried that Harvard would
also make every attempt to meet Pauling’s needs and interests in order to draw him away
from California.519 E.B. Wilson, a physical chemist at Berkeley, sympathized with
516 Letter from Gregory P. Baxter, Chairman of Division of Chemistry, Harvard University, to Pauling, February 2, 1929. LP Safe 2.003. LPP. 517 Pauling, Interview with John L. Greenberg, May 10, 1984, 26. Both Lewis and one of Pauling’s Caltech colleagues, Richard Tolman, had previously turned down offers for the same position. 518 Baxter to Pauling, February 2, 1929. 519 Noyes to Pauling, July 8, 1927.
177
Noyes, telling him, “I suppose that they could much better afford to lose Conant than you
could afford to lose Pauling.”520 Noyes’ fears proved to be not without merit, as Pauling
made serious inquiries to Baxter about more details such as Harvard’s policy on leaves of
absences, the prospect of other appointments in physical chemistry, and the nature of
graduate and undergraduate coursework. As Baxter responded conscientiously to these,
he emphasized the possibility of having him on staff was “a very attractive one”.521 The
entire Chemistry Division was “unanimous in [their] desire” for Pauling to join the staff,
and were “anxious to do everything in [their] power to make [him] feel at home in
Cambridge.”522
Although the reputation and standard of research at Harvard made the offer
undeniably attractive, Pauling’s concerns about whether they could offer a productive
working environment for him were not trivial. He required the installation of and support
for his X-ray and spectrographic apparatus for his experimental research, and the listing
of new graduate courses in chemistry on statistical mechanics, thermodynamics, crystal
structure, and wave mechanics with chemical applications.523 In addition, he requested
the hiring of “[a]t least one other man” to help teach these courses, and was keen to have
“a few men – PhD men and graduate students – to work intimately” with.524, 525 These
requests added up to the greater concern that there would not be a strong community of
physico-chemical researchers to support his work.
520 Letter from E.B. Wilson to Noyes, March 4, 1929, LP Safe 2.003. LPP. 521 Baxter to Pauling, March 7, 1929. 522 Baxter to Pauling, February 26, 1929. 523 Pauling to Conant, March 27, 1929. 524 Pauling to Conant, March 27, 1929. 525 Pauling, Notes on Harvard job offer situation (Undated, March/April/May 1929?), LP Safe 2.003, LPP.
178
Unfortunately Baxter could not be reassuring in this respect, and reluctantly
admitted that the appointment of another professor in physical chemistry was not
something that they could contemplate.526 While Harvard could make every assurance of
supporting Pauling’s individual research needs, the collaborative prospects were less
sure. Although he was attracted to Harvard and its Chemistry division, it was his
foremost wish to “be associated with a department which is strong both in research and
teaching in modern physical chemistry”, and it was not clear that the community would
be there to support it.527
Speculation in the Caltech community proved that Pauling’s concerns about what
environment Harvard could offer were also not unwarranted. Despite its reputation in the
late nineteenth century as a leader in American science, Harvard had by this time
developed a weak reputation in physical chemistry. Wilson was therefore later able to
assuage some of Noyes’ concerns, confiding that “the situation in chemistry at Harvard in
so far as physical chemistry goes is lamentable.”528 He let on that T.W. Richards, in his
tenure, had not been able to teach from more modern texts by Noyes and Sherril or Lewis
and Randall “even with his graduate students except for occasional reference.”529 The
kind of lectures that Pauling was then giving at Berkeley, Wilson believed, “would… be
understood by no member of the Chemical staff at Harvard” with the exception of
students, and then only if they had been “encouraged to pursue mathematics and physics
on the side as they are encouraged [at Berkeley] by Lewis.” Wilson suspected that the
man who was hired in physical chemistry at Harvard would have to build fresh, from the
526 Baxter to Pauling, March 7, 1929. 527 Letter from Pauling to James B. Conant, Harvard Chemistry Division, March 27, 1929, LP Safe 2.003, LPP. Emphasis added. 528 Wilson to Noyes, March 4, 1929. 529 Wilson to Noyes, March 4, 1929.
179
ground up in order to break from their very classical, conservative curriculum. “But
maybe they have decided to make a violent break with their traditions”, Wilson told
Noyes, “and jump from prehistoric into futuristic physical chemistry at one bound!”530
Harvard faculty made every possible reassurance about the quality of the
interdisciplinary connection between physics and chemistry. Baxter had earlier
emphasized that one of the “most attractive features in Cambridge” was “the strong
cooperative feeling… not only inside our Division but also between different
departments.”531 These sentiments were later echoed in greater detail by other chemists,
such as Conant and Edwin C. Kemble. Conant hoped Pauling would join their group and
“talk with [them] about the development of modern physical chemistry at Harvard”.
Everyone,” he wrote, “feels that modern physical chemistry at Harvard should be
developed and should include all that you specify.”532 In the same week, Kemble wrote
in a similar vein. Like Conant, Kemble admitted to weaknesses of the Harvard program,
but was determined to show Pauling that Harvard’s interdisciplinary qualities were not as
frail as he had thought. “It seems to me that you are needlessly pessimistic about your
prospects here”, Kemble told him; “the relations between chemists and physicists are
very cordial.”533
Much of the discussion related to how the work of the Chemistry and Physics
Divisions could combine without creating redundancy, and how connections could be
formed between them. Chemists at Harvard realized that work of Pauling’s kind could
not be done without a proper background in mathematics and physics, and Kemble
530 Wilson to Noyes, March 4, 1929. 531 Baxter to Pauling, February 2, 1929. 532 Conant to Pauling, April 4, 1929. 533 Letter from Edwin C. Kemble, Harvard Physics Division, to Pauling, April 6, 1929, LP Safe 2.003, LPP.
180
suggested to Pauling that the type of students he desired would come forward when the
new programs were created. It was already common enough for chemists to enrol in
physics classes there, and Kemble suggested the physicists would be “glad to have more
migration from physics into chemistry,” especially if new interdisciplinary courses were
to be created.534 Harvard chemists believed creating a whole program in chemistry
around quantum mechanics would be “needless duplication of effort” of what was
already available in Physics, and although the Chemistry Division had been regrettably
“backward” in taking it up for research, many of the faculty there had been “much
interested” in the new research, and the department was “anxious to bring itself abreast
with the times in this direction.”535
As Pauling’s negotiations continued and the importance of a strong disciplinary
connection became clearer to the Chemistry Division, members of Harvard’s physics
community joined in making the pitch. Slater was a particularly enthusiastic exponent of
the physico-chemical atmosphere there, and wrote a lengthy letter in advance of
Pauling’s visit with details about the scientific climate at Harvard and quality of life in
Cambridge. Slater found it hard to believe that he could find California to be as
academically stimulating as the American East Coast:
After all, the whole Los Angeles region tends to sum (sic) to rich retired famous… and while they build delightful homes, and while there is lots of money floating around, I can’t believe that the atmosphere can be so intellectually satisfying as here. I don’t know of any place any where there is such a large number proportionally of really intelligent people as here, with the sort of ideas that one is really interested in.536
534 Kemble to Pauling, April 6, 1929. 535 Kemble to Pauling, April 6, 1929. 536 Letter from Slater to Pauling, May 7, 1929, LP Safe 2.003, LPP.
181
In playing up Harvard’s strengths, Slater also warned Pauling about Caltech’s
weaknesses. He believed there was an “element of danger in the situation at Pasadena”,
because of what Slater perceived as lack of interest by physicists in Pauling’s work;
although chemistry was turning, through quantum mechanics, towards physics, “a really
useful physics department” now had to turn toward chemistry.537 Slater believed Harvard
physicists, inspired by the changing tide of research, were willing to do this.
In Slater’s view, Harvard’s conservative attitude was its greatest strength, because
of its focus on teaching scientific fundamentals. The perspective he offered Pauling was
that “no place in the country” could do better with giving students a “real understanding”
of “modern physics, and of how to apply it”; at Harvard there was “a broad and solid
foundation” on which it was “safe to build as high as possible.”538 The correspondence
between Slater and Pauling gives a strong impression of untapped community potential:
one can only imagine the consequences of a scenario in which these two men could
collaborate in a re-structured department of chemical physics. With Harvard’s existing
strengths, there was undoubtedly the potential for a new physico-chemical research
program to develop quickly into a community, especially with Slater and Pauling at the
same university. “I have a feeling here we could build things up, with comparatively little
effort”, Slater told Pauling, “to be not only the best place in the country, but one of the
very best in the world, in the application of quantum theory.”539 “We surely hope you’ll
come”, he urged him: “I’ve never seen the whole place as unanimously enthusiastic as it
is about the possibility.”
537 Slater to Pauling, May 7, 1929. 538 Slater to Pauling, May 7, 1929. 539 Slater to Pauling, May 7, 1929.
182
However encouraging the Harvard faculty were about the relationship between
their chemists and physicists, what is clear from their correspondence with Pauling is that
the kind of interdisciplinary environment he needed for his work would have to be built
around him upon his arrival. Unlike Caltech, where physicists, chemists and
mathematicians moved freely in the same quantum mechanics seminars and Noyes and
Millikan made interdisciplinary connections a priority, Harvard chemistry was only just
beginning to make the changes that would give such research the framework of an
institutionally designed program. Although Conant, Kemble and Slater were all
encouraging about the personal attitude of the scientists there, this was coupled with
discussion of projects in progress, and promises of what Harvard could accomplish “if”
Pauling was there.
Pauling deliberated quite thoroughly on the Harvard offer. Having “considerable
respect for the solidity and permanence of Harvard” and the “many clever students” who
were attracted there, he did not want to make a decision without considering what both
Caltech and Harvard would also have to offer.540 In his personal notes about the situation,
he counted “[i]ndependence and definiteness” among the strengths of the Harvard
position, along with the “possibility of a coordinated attack on problems of molecular and
crystal structure” with men like Slater.541 Pauling was also intrigued by the possible
benefits from being in a “different sort of country” and the growth that could come from
the new experience.542 At this time, however, he was not keen to “get deeply in to any
field” other than those he was now in, and wanted an environment where he could keep
540 Letter from Pauling to Noyes, February 27, 1929. LP Safe 2.003, LPP. 541 Pauling, Notes on Harvard job offer situation. 542 Pauling, Notes on Harvard job offer situation.
183
experimental crystal structure, quantum mechanics, and molecular structure as the focus
of his research.
The value Harvard placed on his work gave Pauling considerable leverage with
the Caltech administration, and more confidence about identifying his research needs.
Noyes was very interested in providing him with what he needed to stay in California,
and as much as Pauling was concerned that he look out for his professional needs, Noyes’
advice as his former supervisor held considerable personal value.543 At the top of
Pauling’s priorities were travel to other scientific centres, such as Berkeley in particular,
and a dedicated graduate assistant for his work on crystal structure.544 Noyes responded
with confirmation that his interests would be served at Caltech. Although Pauling’s salary
was then only $4500 per year, it would increase to $5000 in the following year. In
addition, he would have a travel grant of $1000 in 1931 for visiting European scientific
centres, and an eventual appropriation of $2000 per year to fund graduate research
assistants.545 Knowing Pauling’s arrangement at Berkeley was a large incentive for him
to stay at Caltech, Noyes offered him a half year leave of absence every two years to
facilitate such visits, something that Harvard would certainly not be able to match.546
In May of 1929, Pauling settled on remaining at Caltech; the Harvard situation
appealed to him, but in the end his position at Caltech and its proximity to Berkeley was
too attractive to give up. Despite Slater’s letter and a further telegram from him
promising “active cooperation” with the Physics Division, which were “effective in
543 Pauling to Noyes, February 27, 1929. 544 Letter from Pauling to Noyes, March 21, 1929, LP Safe 2.003, LPP. 545 Letter from Noyes to Pauling 546 Conant to Pauling, April 6, 1929.
184
upsetting [his] peace of mind”, Pauling was content in his decision to stay.547, 548
Harvard did not take the news without a fight. Desperate not to lose the opportunity
Pauling presented, Conant responded with an urgent telegram two days after Pauling’s
rejection asking him to reconsider, bettering Baxter’s first offer on his “own personal
initiative”: the counter offer promised “a bare possibility” of a full professorship at $6000
salary, and hope of appointing a second man in Pauling’s field in the wake of fresh news
of an opening in Chemistry. Pauling was swayed enough to consult with Millikan, who
further advised him to remain in Caltech, calling the offer of a full professorship at age
twenty-eight an “abnormal honor.”549 “Looking at your interests alone”, Millikan wrote
via telegram, “I advise rejecting Conant’s suggestion”: “Pasadena and Berkeley contacts
together much better”.550
The Harvard offer had worked to Pauling’s advantage by improving his situation
at Caltech, but according to later statements he did not see it as only a bargaining chip; he
simply turned down one job in favour of another that better suited his needs.551 The
greatest loss in the process was in losing the potential of a connection with Slater, to
whom he expressed his regret about having to give up “this cooperative idea”.552 It was
clear that Pauling believed Caltech would be a better home for his research because of
Noyes’ assurance that he would have enough incentive to stay. “If I had only been
dissatisfied with the Institute,” Pauling wrote to Conant in explanation, “everything
would have been different, but I have had almost no cause for complaint.”553 Although
547 Western Union Telegram from Slater to Pauling, May 12, 1929, LP Safe 2.003, LPP. 548 Letter from Pauling to Slater (draft), May 18, 1929, LP Safe 2.003, LPP. 549 Western Union Telegram from Robert Millikan to Pauling, May 21, 1929, LP Safe 2.003, LPP. 550 Millikan to Pauling, May 21, 1929. 551 Pauling, Interview with John L. Greenberg, May 10, 1984. 552 Pauling to Slater, May 18, 1929. 553 Letter from Pauling to Conant (draft), May 18, 1929, LP Safe 2.003, LPP
185
Harvard offered hopes of building a new interdisciplinary program around Pauling’s
work, he still felt that at Caltech he would have “more funds and perhaps more freedom
for research”, and enough staff in physical chemistry to give Pauling the room he needed
to lecture on advanced topics instead of only elementary material.554 In Pauling’s mind
there was “no doubt that the people in charge” at Caltech wanted to make it “ideal for
research”, and were “willing to do whatever helps toward that.”555
Slater was disappointed not to secure Pauling as part of the Harvard community,
but did not give up on the idea of working with him in the same department. A year after
the Harvard offer, Slater moved to M.I.T. on the invitation of Karl Compton, to direct the
physics department there. In the years since Noyes’ departure, science programs at
M.I.T. had fallen into a comparative decline, and Compton had his eye on Slater to
revitalize them. Extending his earlier goal of developing the quality of American science,
Slater believed the best way to carry this out at M.I.T. was to form stronger
interdisciplinary programs, and closer connections with Harvard. When the news of his
former academic home reached Noyes, he was “delighted to hear” that M.I.T. was now
“likely to become a great research center of physics” under Slater’s care, and hoped that
chemistry would soon follow.556 Slater knew it would be “a hard job” to convince
Pauling to move, but was determined to do so, the prospect a joint chemical physics
initiative being too good to ignore.557
In January of 1931 Slater wrote to Pauling with an offer of a full professorship at
M.I.T. The offer was impressive, and tailored to Pauling’s specific needs: a yearly salary
554 Pauling to Conant, May 18, 1929. 555 Pauling to Slater, May 18, 1929. 556 Letter from Noyes to Pauling, December 23, 1930, LP Safe 2.017, LPP. 557 Letter from Slater to Pauling, January 21, 1931, LP Safe 2.003, LPP.
186
of $8000 with the possibility of a joint appointment in physics and chemistry. He pitched
M.I.T. to Pauling with much the same enthusiasm as he had done for Harvard in 1929,
and spoke as positively, if not more so, about the prospects for Pauling because of the
combined strength of Harvard and M.I.T.’s physico-chemical resources. He told Pauling
enthusiastically that between the two Cambridge schools, “there is no question but we
can build up a scientific group as good as any in the country”, with a particular value for
the emerging chemical physics.558 There were plans to construct a new laboratory with,
as Slater described it, “the very fine and almost unique feature” of being “a combined
research laboratory for physics and chemistry.”559
Responding to Pauling’s earlier fears of conservatism and research isolation at
Harvard, Slater reassured him that he would have no such concerns with M.I.T. “I am
finding that I work with the chemists a great deal,” he remarked, “and they are very co-
operative, and well informed on modern physics. now there is enough foundation both
here and at Harvard, so that you would have no fear of having to carry on by yourself.560
He further boasted that Cambridge was on its way to becoming “the scientific center of
the country” with the strength of the team Slater was assembling at M.I.T.561 He now felt
that, with both Harvard and M.I.T. moving forward along progressive research lines,
Cambridge offered a combination that could not even be duplicated by the situation
Pauling enjoyed at Pasadena. “Don’t you begin to admit,” he nudged his west-coast
colleage, “that Cambridge is an interesting place to be?”562
558 Letter from Slater to Pauling, June 25, 1930, LP Safe 2.003, LPP. 559 Slater to Pauling, June 25, 1930. 560 Slater to Pauling, January 21, 1931. 561 Slater to Pauling, January 21, 1931. 562 Slater to Pauling, June 25, 1930.
187
Pauling considered the M.I.T. offer carefully but as he had decided after the
previous offer from Harvard, he was too satisfied with his position at Caltech to accept.
As he expressed his ambivalence to Noyes about the situation, he admitted that he had
“felt for some time” that he and Slater “could very profitably work together,” and told
Slater as much when he wrote with the disappointing news.563, 564 “I have come to
realize,” he wrote his colleague, “that there is no theoretical physicist whose work
interests me more than yours.”565 Pauling doubted, though, that M.I.T. would be able to
provide research facilities that could match Caltech’s. Although he regretted missing the
opportunity to work with Slater on the “structural problems” which interested them both
and the contact with Cambridge physicists, he was very happily satisfied to remain part
of physical chemistry in Pasadena.
It was also clear that in addition to the attractiveness of Caltech, the connection
with Berkeley was what tipped the scales in Pauling’s decision to remain in California. In
response to Conant and Slater’s eleventh-hour missives he explained that the lectures at
Berkeley attracted him especially: he “weighed everything as carefully as [he] could, and
tried hard to decide to come East, but it was of no use.”566 The possibility of lecturing on
new material in what he hoped would be an annual visit with Lewis’ group “had some
influence” on his decision, coupled with the opportunity for absences every two years.567
The feeling was certainly mutual. Once Lewis learned of the decision, he wrote Pauling
on behalf of his entire group to say that they were “all pleased” to know he was “not
563 Ever the politician, Noyes responded to word of Slater’s offer with the question of whether they should invite Slater to Caltech for a series of guest lectures and collaboration with Pauling. Letter from Noyes to Pauling, May 13, 1930. 564 Letter from Pauling to Noyes, January 27, 1931, LP Safe 2.003, LPP. 565 Letter from Pauling to Slater, February 9, 1931, LP Safe 2.003, LPP. 566 Pauling to Slater, May 18, 1929. 567 Pauling to Conant, May 18, 1929.
188
going to be far away.”568 “It is a delight,” Lewis wrote, “to know that you can be with us
and almost any time that suits you will suit us.”569
The combined community of Caltech and Berkeley made an ideal environment
for Pauling’s research. In Noyes’ group he had the benefits of working in a small, elite
research institution led by some of the most influential American scientists of the early
twentieth century, a place where physical chemistry was developing rapidly under Noyes’
direction. At Berkeley, only a few hours’ drive away, Pauling had an opportunity to
develop his research in two important ways. First, he was permitted to lecture on topics
related to his specific field, rather than the general physical chemistry teaching that would
have been required of him at Harvard or M.I.T. Second, Lewis being the head of the
chemistry department made Berkeley an ideal place for Pauling to bring his research on
the chemical bond. Although Harvard and M.I.T. were respected institutions that offered
high-calibre physico-chemical research programmes, the connection with Lewis, the
founder of valence bond theory was part of what made Caltech special to Pauling.
5. Chemical bond in context: reductionism, interdisciplinary, and the path of disciplinary change
From the late nineteenth century through to the 1930s, physico-chemical research
grew to encompass an ever widening field of problems and academic homelands. What
began as a well-defined institutional movement to formalize the use of physical methods
in electrolytic and thermodynamic chemical analysis gave way, half a century later, to an
entirely different disciplinary framework where chemists and physicists contributed to
problems of collaborative value in the wake of new physical discoveries. It was difficult 568 Letter from Lewis to Pauling, May 29, 1929. LP Safe 3.018, LPP. 569 Lewis to Pauling, May 29, 1929.
189
for the chemical bond to find a place in the older physical chemistry, where atomistic
thinking was constrained by first the positivist concerns of its founders and next, the
limitations of chemical methods. Applying wave mechanics to the study of electron-
sharing made chemical bond research a broader, sophisticated and more challenging area
of study. This research also let in new physical tools to chemistry and showed physicists
the potential in problems from a different discipline. In this emergence of chemical
physics, the chemical bond can be studied as a signifier of disciplinary change, that
allows us to identify with the experiences and issues that new chemical physicists like
Pauling faced.
Unlike its disciplinary predecessor, physical chemistry, which could be clearly
defined by its institutional homeland of the University of Leipzig, and electrolytic and
thermodynamic analyses of its founders the Ionists, chemical physics must be defined
much more amorphously across a range of institutions and problems of study. During the
earliest years of Pauling’s career, over the late 1920s and early 1930s, as he was
challenged to discover his professional identity, he also faced the question of where his
work would be valued. In his story, his disciplinary homeland was the community that
made his work possible, and this was not limited to one side of the physico-chemical
border. Chemical physics also has its signposts, most notably the creation of the Journal
of Chemical Physics in 1933, but the real moments of change are behind the scenes, in
the daily trappings of academia in a period of emergence. Here, what defined Pauling as a
chemical physicist was not a personal desire to re-write the physico-chemical disciplinary
landscape, but the realization in the physico-chemical community that this landscape
needed to be changed in order to accommodate research like his.
190
These fruitful issues of interdisciplinarity that make the chemical bond such a
clear case study for the disciplinary convergence of chemistry and physics make clear
that reductionist arguments fail to capture the chemical bond’s true historical and
philosophical significance. The argument that the Heitler-London treatment and its
ensuing foundational importance within quantum chemistry allows chemistry to be
theoretically reduced to physics does not allow us to understand its significance as a
theoretical entity that was shaped by physical and chemical research. At the heart of any
problem in chemistry is the combination of atoms to form molecules, but since chemists
are unable to describe the mechanics of the electron or explain the physical basis of its
ability to bond through shared pairs, some degree of physical intervention is necessary to
study the chemical bond. Whether or not we consider reductionism to be an accurate
consequence of the Heitler-London treatment, there is no avoiding the fact that until
physical methods were available to chemists, there was no real possibility of discovering
the mechanism of bonding.
What is also evident here is that the interdisciplinarity of the chemical bond
manifests in two distinct ways: first, the nature of the theory itself as a physico-chemical
synthesis, and second, the mixed academic climate that fostered its development. Pauling,
Langmuir, Slater and Mulliken all produced versions of a physico-chemical theory,
according to their disciplinary influences and backgrounds. Pauling’s experiences in his
early career further exemplify the challenges in classifying chemical bond research in the
post-quantum mechanics era, and strengthen the historical contrast between chemical
physics and physical chemistry. Unlike physical chemists who have had the comparative
luxury of a set of founding fathers who laid goals and a path of institutional growth for
191
their discipline, chemical physicists forged their discipline more amorphously from
independent, autonomous contributors.
Pauling’s experiences also raise questions about the path taken towards
disciplinary change, in the broader context of physical sciences in the early 20th century.
There was no ideological crusade in his pursuit of interdisciplinary research, and he set
no specific goals in combining physics and chemistry, but the nature of his research
required him to make connections outside the physical chemistry community of the early
1920s. His interdisciplinary work required interdisciplinary resources, and so the
chemical bond itself was what led him to the emerging areas of chemical physics and
quantum chemistry. His process of discovery led to outcomes only natural for a junior
scientist working to carve out an area of research that set him apart from his predecessors.
He was lucky enough to find this niche in the trend-setting area of quantum chemistry,
but it was not luck that led him to the valuable communities of Munich, Caltech, and
Berkeley: these were the places where his work was accepted, and was valued. In the
early years of the late 1920s and early 1930s, the major questions Pauling needed to ask
were of how to define his work in his homeland of chemistry. As it became clear that his
field was not physics, not chemistry proper, and not classical physical chemistry, it was
then that a different disciplinary landscape began to emerge. By incorporating methods
from diverse traditions, he made himself open to diverse communities.
It is rather the normalcy of Pauling’s experiences that is most striking. Here we
find very recognizable elements of academia: the guest lecture, the department colloquia,
the hiring committee, and the project supervisor. The Ionists had worked deliberately to
form political connections in the 1880s, to self-identify within the community as part of a
192
new scientific practice. This was a necessary part of making physical chemistry
recognizably distinct, along with the formation of the Zeitschrift and Ostwald’s Leipzing
homeland. Pauling’s experiences were at a basic level, natural steps taken by a young
academic. Given the contrasts between the more fluid growth in American universities,
and the more rigid institutional structures of German science, the chemical bond story
suggests disciplinary emergence is influenced by national styles and scientific
frameworks. In broader historical terms, it raises questions of disciplinary development in
the twentieth century, given the even more amorphous boundaries exhibited by such
areas as molecular biology, which grew from the chemical bond’s successes. When so
many fragmented traditions make up such a broad whole, it can only be risky to define
chemical physics too closely.
In the terms of the bond as a theoretical work, interdisciplinarity prevents us from
adopting a reductionist viewpoint, because we must acknowledge the parts physics and
chemistry play in its solution. Chemical autonomy helps us understand why, on a
theoretical level, chemists relied on a chemically-motivated theory of valence instead of a
physical theory of the atom, and how chemists and physicists could approach electrons
and atoms from different perspectives. Unlike Scerri’s suggestion that adopting a
chemically autonomous viewpoint on physical entities with chemical value makes it safe
for chemists to think differently about them, autonomy must instead be acknowledged on
both sides because of the historical nature of how the theory developed. While atomic
orbitals may be given by pure physics, the chemical bond as a whole cannot. The
chemical bond forged a connection between the two disciplines, and this theoretical
convergence conserved Lewis’ original theory of valency, while also relying on physical
193
theory to explain the mechanism of chemical behaviour. The physical and chemical
contributions to the theory itself play a direct role in determining the nature of chemical
physics because of the value chemical physicists find in studying it.
It is this emergence of a blend of worldviews and the problem of disciplinary
ownership that must be emphasised when approaching the chemical bond. We need a
way to treat the complexity of the historical and philosophical issues that the chemical
bond presents, of theory reduction, disciplinary synthesis, disciplinary reduction,
theoretical synthesis, and the ontological status of interdisciplinary theories. Rather than
defend the respective claims of physics and chemistry to this problem, we must
acknowledge both claims and allow the chemical bond to be regarded as a physico-
chemical entity. We cannot ignore the role of physics in the case of the valence bond; but
because of the contributions of Lewis and Pauling, chemistry will always claim some part
of its ownership. Because physics and chemistry have each had a share in its
development we should not allow claims from only one discipline to motivate our
analysis of this case.
Science in a borderland is problematic, because it cannot be easily categorized, or
owned by only one discipline. At the simplest level we may regard the valence bond as a
work that, historically, forced chemists and physicists to confront the boundaries of their
disciplines; at another level we may treat it as the centre of a complex convergence of
issues. For nineteenth century chemists the chemical bond, like the atom and molecule,
was a conceptual entity that could not be studied by purely chemical means. But through
the convergence of physical and chemical research in the twentieth century, chemists
gained the physical and theoretical tools they needed to be able to study bonds. In the
194
1920s and 1930s, both physicists and chemists sought to identify with the chemical bond
because it opened up new challenges and research questions for their respective fields. It
is exciting to consider how quantum mechanics changed the boundaries of physical and
chemical work because of the problem of the chemical bond, and it is the nature and
cause of this connection that should be studied.
The key issue for the chemical bond is less one of theoretical reduction than one
of disciplinary ownership and theoretical convergence. At the simplest level we may
regard the valence bond simply as a theory that, historically, forced chemists and
physicists to confront the boundaries of their disciplines; at another level we may treat it
as the centre of a complex convergence of issues. In the 1920s and 1930s, both physicists
and chemists sought to identify with the chemical bond because it opened up new
challenges and research questions for their respective fields. As a chemical problem that
cannot be explained without physical methods, it is something to which neither physicists
nor chemists can fully stake a claim.
195
Bibliography
Barkan, Diana, Walther Nernst and the transition to modern physical science (Cambridge:
Cambridge University Press, 1999).
Beller, Mara, Quantum Dialogue: The Making of a Revolution (Chicago: University of
Chicago Press 1999).
Bensaude-Vincent, Bernadette (ed.), Communicating Chemistry: Textbooks and their
Audiences (Canton, Massachussetts: Science History Publications, 2000).
Born, Max, The Constitution of Matter (London: Metheun, 1923).
Bragg, W.H., and Bragg, W.L., X-Rays and Crystal Structure (London: G. Bell and Sons,
1915).
Bray, William C., and Branch, Gerald E. K., “Valence and Tautomerism”, Journal of the
American Chemical Society, 35 (1913), 1440-1447.
Branch, Gerald E. K., “Appendix: Gilbert Newton Lewis, 1875-1946”, Journal of
Chemical Education, 61 (1984), 18-21, Appendix to Melvin Calvin, “Gilbert Newton
Lewis: His Influence on Physical-Organic Chemists at Berkeley,” same issue, 14-18.
Brock, W.H., The Atomic Debates: Brodie and the Rejection of the Atomic Theory
(Leicester University Press, 1967).
Brock, W.H, The Norton History of Chemistry (New York: W.W. Norton & Co, 1993).
Brodie, Benjamin C., “The Calculus of Chemical operations; being a Method for the
Investigation, by means of Symbols, of the Laws of the Distribution of Weight in
Chemical Change. – Part I. On the Construction of Chemical Symbols,” Philsophical
Transactions of the Royal Society of London, Vol 156 (1866): 781-859.
196
Brush, Stephen G. “Dynamics of Theory Change in Chemistry. Part 1: The Benzene
Problem, 1865-1945”, Studies in the History and Philosophy of Science, 30 (1999),
21-79.
Brush, Stephen G. “Dynamics of Theory Change in Chemistry. Part 2: Benzene and
Molecular Orbitals, 1945-1980.”, Studies in the History and Philosophy of Science,
30, (1999), 263-302.
Crawford, Elisabeth, “Arrhenius, the Atomic Hypothesis, and the 1908 Nobel Prizes in
Physics and Chemistry,” Isis, 75 (1984), 503-522.
Darrigol, Oliver, From c-numbers to q-numbers : the classical analogy in the history of
quantum theory (Berkeley: University of California Press, 1992).
Daston, Lorraine (ed.), Biographies of Scientific Objects (Chicago: University of Chicago
Press, 2000).
Davenport, Derek A., “In This Issue,” Journal of Chemical Education, 61 (1984), 2.
Douglas Clark, C.H., “Some Physical Aspects of Atomic Linkages,” Chemical Reviews
11 (1932), 231-271.
Dunitz, Jack D. “Linus Carl Pauling: 20 February 1901 – 19 August 1994”, Biographical
Memoirs of the Fellows of the Royal Society, 42 (1996), 316-338.
Ewald, P.P. (ed.), Fifty Years of X-Ray Diffraction (International Union of
Crystallography: Utrecht, 1962).
Gasman, L, “Myths and X-rays,” The British Journal for the Philosophy of Science, 26
(1975), 51-60.
197
Gavroglu, Kostas, “The Physicists’ Electron and its Appropriation by the Chemists”, in
Histories of the Electron, Buchwald, Jed and Warwick, Andrew (editors) (Cambridge:
M.I.T. Press, 2001), 363-400.
Gavroglu, Kostas, and Simões, Ana, “The Americans, the Germans, and the beginnings
of quantum chemistry: the confluence of diverging traditions.” Historical Studies in
the Physical Sciences, 25 (1994), 47-110.
Gavroglu, Kostas, and Simões, Ana. “Preparing the ground for quantum chemistry in
Great Britain: the work of the physicist R.H. Fowler and the chemist N.V. Sidgwick”,
British Journal for the History of Science, 35 (2000), 187-212.
Gavroglu, Kostas, and Simões, Ana, “One Face or Many: The Role of Textbooks in
Building the New Discipline of Quantum Chemistry”, in Bensaude-Vincent, Bernadette
(ed.), Communicating Chemistry: Textbooks and their Audiences 1789-1939 (Canton,
Mass.: Science History Publications/USA, 2000), 415-449.
Geiger, Roger L., To Advance Knowledge: The Growth of American Research Universities,
1900-1940 (New York: Oxford University Press, 1986).
Gibbs, J. Willard, Scientific Papers (Dover Publications: New York, 1961).
Gillespie, R.J., and Robinson, E.A., “Gilbert N. Lewis and the Chemical Bond: The
Electron Pair and the Octet Rule from 1916 to the Present Day”, Journal of
Computational Chemistry, 28 (2007), 87-97.
Goertzel, Ted, and Goertzel, Ben, Linus Pauling: A Life in Science and Politics (New
York, NY: BasicBooks, Harper Collins, 1995).
Judith R. Goodstein, Millikan’s school: a history of the California Institute of
Technology (W.W. Norton, 1991).
198
Goupil, Michelle, Du Flou au Clair (Editions du comité des travaux historiques et
scientifiques: Paris, 1991).
Hager, Thomas, Force of Nature: The Life of Linus Pauling (New York: Simon &
Schuster, 1995).
Hawkins, Laurence Ashley, Adventure Into the Unknown: The first fifty years of the
General Electric Research Laboratory (Morrow Press, 1950).
Heilbron, John L., “The Kossel-Sommerfeld Theory and the Ring Atom”, Isis, 58 (1967),
450-485.
Heitler, Walter and London, Fritz, “Wechselwirkung neutraler Atome und homopolare
Bindung nach der Quantenmechanik”, Zeitschrift fur Physik, 44 (1927), 455-472.
Hendry, Robin Findlay, “The Physicists, the Chemists, and the Pragmatics of Explanation”,
Philosophy of Science, 71 (2004), 1048-1059.
Hiebert, Erwin N., and Korber, Hans-Gunther, “Wilhelm Friedrich Ostwald”, Dictionary of
Scientific Biography (New York: Scribner, 1981, c1980-c1990), Vol 15, Supplement I,
455-469.
Hildebrand, Joel, “Gilbert Newton Lewis: October 25, 1875-March 23, 1946”,
Biographical Memoirs, National Academy of Sciences, Vol 31 (1947), 210-235.
Janich, Peter and Psarros, Nikolaos, The Autonomy of Chemistry: 3rd Erlenmeyer-
Colloquy for the Philosophy of Chemistry (Würzburg, Königshausen and Neuman,
1998).
Jenkins, Zach, “Do You Need to Believe in Orbitals to Use Them?: Realism and the
Autonomy of Chemistry.” Philosophy of Science, 70 (2003), 1052-1062.
199
Jungnickel, Christa and McCormmach, Russell, Intellectual Mastery of Nature:
Theoretical Physics from Ohm to Einstein (2 vol.) (University of Chicago Press,
1986).
Kaiser, David, Pedagogy and the practice of science: historical and contemporary
perspectives (Cambridge: M.I.T. Press, 2005).
Karachalios, Andreas, “On the Making of Quantum Chemistry in Germany”. Studies in
the History and Philosophy of Modern Physics, 31 (2000), 493-510.
Kevles, Daniel, “The Physics, Mathematics, and Chemistry Communities: A Comparative
Analysis”, in Oleson and Ross, eds, The Organization of Knowledge in Modern
America, 1860-1920, 139-172.
Kevles, Daniel, The Physicists (fourth edition) (Harvard University Press, 1995).
Kim, Mi Gyung, Affinity, That Elusive Dream: A Genealogy of the Chemical Revolution
(Cambridge: M.I.T. Press, 2003).
Kipnis, Alexander Y., “Early Chemical Thermodynamics: Its Duality Embodied in Van’t
Hoff and Gibbs”, in W. Hornix and S.H.W.M. Mannaerts (editors), Van’t Hoff and the
Emergence of Chemical Thermodynamics: Centennial of the first Nobel Prize for
Chemistry 1901-2001 (Delft: D.U.P. Science, 2001), 212-242.
Kistiakowsky, G.B. “Review: Pauling, Linus, The nature of the chemical bond and the
structure of molecules and crystals (Cornell University Press, 1939).”, Journal of the
American Chemical Society, 51 (1939), 457.
Klein, Ursula, Experiments, Models, Paper Tools: Cultures of Organic Chemistry in the
Nineteenth Century (Stanford University Press, 2003).
200
Kohler, Robert E, “The Origin of G.N. Lewis’s Theory of the Shared Pair Bond,”
Historical Studies in the Physical Sciences, 3 (1971), 343-376.
Kohler, Robert E., “Lewis, Gilbert Newton”, Dictionary of Scientific Biography, Vol 8
(Charles Scribners & Sons: American Council of Learned Societies, Charles Coulson
Gillispie, ed., 1973), 289-293.
Kohler, Robert E., “Irving Langmuir and the Octet Theory of Valence,” Historical
Studies in the Physical Sciences, 4 (1974), 39-87.
Kohler, Robert E., “The Lewis-Langmuir Theory of Valence and the Chemical
Community, 1920-1928”, Historical Studies in the Physical Sciences, 6 (1975), 431-
468.
Kohler, Robert E., “The Ph.D. Machine: Building on the Collegiate Base”, Isis, 81 (1990),
638-662.
Kragh, Helge, Quantum generations: a history of physics in the twentieth century
(Princeton: Princeton University Press, 1999).
Kragh, Helge, “Van’t Hoff and the Transition from Thermochemistry to Chemical
Thermodynamics”, in W. Hornix and S.H.W.M. Mannaerts (editors), Van’t Hoff and
the Emergence of Chemical Thermodynamics: Centennial of the first Nobel Prize for
Chemistry 1901-2001 (Delft: D.U.P. Science, 2001), 191-211.
Kuhn, Thomas S., Black-body Theory and the Quantum Discontinuity 1894-1912
(Oxford University Press, 1978).
Lachman, Arthur, Borderland of the Unknown: The Life Story of Gilbert Newton Lewis
(Pageant Press, 1955).
201
Laidler, Keith J., The World of Physical Chemistry (Oxford: Oxford University Press,
1995).
Laidler, Keith J., “Van’t Hoff and the Birth of Chemical Dynamics”, in W. Hornix and
S.H.W.M. Mannaerts (editors), Van’t Hoff and the Emergence of Chemical
Thermodynamics: Centennial of the first Nobel Prize for Chemistry 1901-2001 (Delft:
D.U.P. Science, 2001), 243-255.
Langmuir, Irving, “The Arrangement of Electrons in Atoms and Molecules”, Journal of
the American Chemical Society, 41 (1919), 868-934.
Langmuir, Irving, “Isomorphism, Isosterism and Covalence”, Journal of the American
Chemical Society, 41 (1919), 1543-1559.
Langmuir, Irving, “Modern Concepts in Physics and Their Relation to Chemistry,”
Science, 70 (Oct. 25, 1929), 385-396.
Von Laue, M, Friedrich,W. and Knipping, P, “Interferenz-Erscheinungen bei
Rontgenstrahlen,” Sitzungsberichte der Bayerische Akademie der Wissenschaften:
physikalische-mathematische klasse, 42 (1912), 303-324.
Levere, Trevor, Affinity and Matter: Elements of Chemical Philosophy 1800-1865
(Oxford: Clarendon Press, 1971).
Lewis, Gilbert N., “Valence and Tautomerism,” Journal of the American Chemical
Society, 35 (1913), 1448-1455.
Lewis, Gilbert N., “The Atom and the Molecule,” Journal of the American Chemical
Society, 38 (1916), 762-785.
Lewis, Gilbert N., “The Static Atom”, Science, 46 (Sep 28, 1917), 297-302.
202
Lewis, Gilbert N., Valence and the Structure of Atoms and Molecules (American
Chemical Society Monograph Series: The Chemical Catalog Company, New York,
1923).
Lewis, Gilbert N., “The Magnetochemical Theory,” Chemical Reviews, 1 (1925), 231-
245.
Massimi, Michela, Pauli’s Exclusion Principle: the origin and validation of a scientific
principle (Cambridge University Press, 2005).
Merz, John Thedore, A History of European Scientific Thought in the Nineteenth Century,
Volume 1 (Dover: New York, 1965) (2 vol reprint).
Millikan, Robert, “Atomism in Modern Physics”, Journal of the Chemical Society, 125
(1924), 1405-1417.
Mott, Nevil, “Walter Heinrich Heitler: 2 January 1904 – 15 November 1981”,
Biographical Memoirs of the Fellows of the Royal Society, 28 (1982), 140-151.
Mulliken, Robert S., “The Assignment of Quantum Numbers for Electrons in
Molecules”, Physical Review 32 (1928), 186-222.
Mulliken, Robert S., “The Assignment of Quantum Numbers for Electrons in Molecules.
III. Diatomic Hydrides”, Physical Review, 33 (1929), 730-747.
Mulliken, Robert S., “The Assignment of Quantum Numbers for Electrons in Molecules.
II. Correlation of Molecular and Atomic Electron States”, Physical Review, 32
(1928), 761-772.
Mulliken, Robert S., “Electronic Structures of Polyatomic Molecules and Valence”,
Physical Review, 40 (1932), 55-62.
203
Mulliken, Robert S., “Electronic Structures of Polyatomic Molecules and Valence. II.
General Considerations”, Physical Review, 41 (1932), 49-71.
Mullikan, Robert S. “Review: Linus Pauling, ‘The nature of the chemical bond’”, The
Journal of Physical Chemistry, 44 (1940), 827-828.
Murrell, John, Kettle, Sydney and Tedder, John, The Chemical Bond (John Wiley and
Sons, 1985).
Nye, Mary Jo, Molecular reality: a perspective on the scientific work of Jean Perrin (New
York: Elsevier, 1972).
Nye, Mary Jo, “The Nineteenth-Century Atomic Debates and the Dilemma of an
‘Indifferent Hypothesis’”, Studies in the History and Philosophy of Science, 7 (1976),
245-268.
Nye, Mary Jo (ed), The Question of the Atom: From the Karlsruhe Congress to the First
Solvay Conference, 1800-1911 (Los Angeles: Tomash, 1984).
Nye, Mary Jo, From Chemical Philosophy to Theoretical Chemistry: Dynamics of Matter
and Dynamics of Disciplines (1800-1950) (University of California Press, 1993).
Ostwald, Wilhelm “Emancipation from Scientific Materialism”, Science Progress, 4 (Feb.
1896), reprinted in Mary Jo Nye (ed), The Question of the Atom: From the Karlsruhe
Congress to the First Solvay Conference, 1800-1911 (Los Angeles: Tomash, 1984), pp
377-354.
Paradowski, Robert, The Structural Chemistry of Linus Pauling (Ph.D. Thesis, University
of Wisconsin, 1972).
Park, Buhm Soon, Computations and Interpretations: The Growth of Quantum
Chemistry, 1927-1967. (Johns Hopkins University: Ph.D. Thesis, 1999).
204
Park, Buhm Soon, “Chemical translators: Pauling, Wheland and their strategies for
teaching the theory of resonance”, British Journal for the History of Science, 32
(1999), 21-46.
Park, Buhm Soon, “The contexts of simultaneous discovery: Slater, Pauling, and the
Origins of Hybridization”, Studies in the History and Philosophy of Modern Physics,
31 (2000), 451-474.
Park, Buhm Soon, “The ‘Hyperbola of Quantum Chemistry’: the Changing Practice and
Identity of a Scientific Discipline in the Early Years of Electronic Digital Computers,
1945-65”, Annals of Science, 60 (2003), 219-247.
Partington, J.R., A History of Chemistry, Volume 4 (London: MacMillan & Co. Ltd, 1964).
Pauling, Linus, “The Dynamic Model of the Chemical Bond and its Application to the
Structure of Benzene”, Journal of the American Chemical Society 48 (1926), 1132-
1143.
Pauling, Linus, “The Theoretical Prediction of the Physical Properties of Many-Electron
Atoms and Ions. Mole Refraction, Diamagnetic Susceptibility, and Extension in
Space”, Proceedings of the Royal Society of London, Series A 104 (1927), 181-211.
Pauling, Linus, “The Sizes of Ions and the Structure of Ionic Crystals”, Journal of the
American Chemical Society, 49 (1927), 765-791.
Pauling, Linus, “The Shared Electron Bond”, Proceedings of the National Academy of
Sciences, 14 (1928), 359-362.
Pauling, Linus, “The Nature of the Chemical Bond. Application of results obtained from
the quantum mechanics and from a theory of paramagnetic susceptibility to the
205
structure of molecules.”, Journal of the American Chemical Society, 53 (1931), 1367-
1400.
Pauling, Linus, “The Nature of the Chemical Bond. II. The One-Electron Bond and the
Three-Electron Bond.”, Journal of the American Chemical Society, 53 (1931), 3225-
3237.
Pauling, Linus, “The Nature of the Chemical Bond. III: The Transition from One
Extreme Bond Type to Another”, Journal of the American Chemical Society, 54(1932),
988-1003.
Pauling, Linus, “The Nature of the Chemical Bond. IV. The Energy of Single Bonds and
the Relative Electronegativity of Atoms”, Journal of the American Chemical Society,
54 (1932), 3570-3582.
Pauling, Linus, The Nature of the Chemical Bond and the Structure of Molecules and
Crystals: An Introduction to Modern Structural Chemistry (Ithaca, New York: Cornell
University Press, 1939).
Pauling, Linus, “Arthur Amos Noyes: September 13, 1866 – June 3, 1936”, Biographical
Memoirs of the National Academy of Sciences 31(1947), 322-346.
Pauling, Linus, and Wheland, G.W., “The Nature of the Chemical Bond. V. The
Quantum-Mechanical Calculation of the Resonance Energy of Benzene and
Naphthalene and the Hydrocarbon Free Radicals.”, Journal of Chemical Physics
1(1933), 362-374.
Psarros, Nikos, “The Lame and the Blind, or How Much Physics Does Chemistry
Need?”, Foundations of Chemistry, 3 (2001), 241-249.
206
Reich, Leonard S., “Irving Langmuir and the Pursuit of Science and Technology in the
Corporate Environment”, Technology and Culture, 24 (1983), 199-221.
Reinhardt, Carsten, “Chemistry in a Physical Mode: Molecular Spectroscopy and the
Emergence of NMR”, Annals of Science, 61 (2004), 1-32.
Rocke, Alan, Chemical Atomism in the Nineteenth Century: From Dalton to Cannizzaro
(Ohio State University Press: Columbus, 1984).
Rodebush, Worth H, “The Electron Theory of Valence”, Chemical Reviews, 5 (1928),
509-531.
Root-Bernstein, Robert S., The Ionists: Founding Physical Chemistry, 1872-1890 (Ph.D.
Thesis, Princeton University, 1980).
Rosenfeld, Albert, “The Quintessence of Irving Langmuir”, The Collected Works of
Irving Langmuir, Volume Twelve (Pergamon Press, Inc, 1962), 3-229.
Russell, Colin A., The History of Valency (Leicester University Press, 1971).
Servos, John W., “G.N. Lewis, The Disciplinary Setting,” Journal of Chemical Education,
61 (1984), 5-10.
Servos, John, “Mathematics and the Physical Sciences in America, 1880-1930,” Isis, 77
(1986), 611-629.
Servos, John W., Physical Chemistry from Ostwald to Pauling: The Making of a Science
in America (Princeton University Press, 1990).
Scerri, Eric R. “Has Chemistry at Least Been Approximately Reduced to Quantum
Mechanics?”, PSA: Proceedings of the Biennial Meeting of the Philosophy of Science
Association, Vol. 1994, 160-170.
207
Scerri, Eric R, “Philosophy of Chemistry: A New Interdisciplinary Field?”, Journal of
Chemical Education, 77 (2000), 522-525.
Schweber, Samuel S., “The Young John Clarke Slater,” Historical Studies in the Physical
Sciences 20 (1990), 339-406.
Shaik, Sason, “The Lewis Legacy: The Chemical Bond – A Territory and Heartland of
Chemistry”, Journal of Computational Chemistry 28 (2007), 51-61.
Shaik, Sason, and Hiberty, Philippe C., “Valence Bond Theory, Its History, Fundamentals,
and Applications: A Primer”, Reviews in Computational Chemistry, 20 (2004), 1-119.
Sherill, Miles S., “The Contributions of Arthur A. Noyes to Science”, Science 84 (1936),
217-220.
Sidgwick, Nevil V., The Electronic Theory of Valency, (Oxford: Oxford University Press,
1927).
Simões, Ana, Converging trajectories, diverging traditions: Chemical bond, valence,
quantum mechanics and chemistry, 1927-193, (University of Maryland, Ph.D. Thesis,
1993).
Simões, Ana, and Gavroglu, Kostas, “Issues in the History of Theoretical and Quantum
Chemistry, 1927-1960”, In Reinhardt (ed.), Chemical Sciences in the 20th Century:
Bridging Boundaries, (Wiley-VCH, 2001).
Simões, Ana, “In Between Worlds: G.N. Lewis, the Shared Pair Bond and Its
Multifarious Contexts”, Journal of Computational Chemistry, 28 (2007), 62-72.
Slater, John C., “The Normal State of Helium”, Physical Review, 32 (1928), 349-360.
Slater, John C., “The Theory of Complex Spectra”, Physical Review, 34 (1929), 1293-
1322.
208
Slater, John C., “Cohesion in Monovalent Metals”, Physical Review, 35 (1930), 509-529.
Slater, John C. “Directed Valence in Polyatomic Molecules”, Physical Review, 37 (1931),
481-489.
Slater, John C., “Molecular Energy Levels and Valence Bonds”, Physical Review, 38
(1931), 1109-1144.
Slater, John C., “Note on Molecular Structure (Letter to the Editor)”, Physical Review, 41
(1932), 255-7.
Slater, John C., Introduction to Chemical Physics (McGraw-Hill, 1933).
Snelders, H.A.M., “Arrhenus, Svante August”, Dictionary of Scientific Biography (New
York: Scribner, 1981, c1980-c1990) Vol 1, p 296-302.
Taylor, Hugh, “Irving Langmuir, 1881-1957”, Biographical Memoirs of the Fellows of
the Royal Society, 4 (1958), 167-184.
Thomson, J.J., “On The Structure of the Atom: an Investigation of the Stability and
Periods of Oscillation of a number of Corpuscles arranged at equal intervals around
the Circumference of a Circle; with Application of the results to the Theory of
Atomic Structure”, Philosophical Magazine, 6 (1904), 239-265.
Thomson, J.J., “On the Structure of the Atom”, Philosophical Magazine, 26 (1913), 792-
799.
Thomson, J.J., “The Forces Between Atoms and Chemical Affinity”, Philosophical
Magazine, 27 (1914), 757-789.
Thomson, J.J., The Electron in Chemistry (Philadelphia: The Franklin Institute, 1923).
209
210
Vance, John E., “Review: Pauling, Linus, The nature of the chemical bond and the
structure of molecules and crystals (Cornell University Press, 1939).”, American
Journal of Science, Dec. 1939, 923-4.
Van der Waerden, Bartel, L., Sources of Quantum Mechanics (Amsterdam: North-
Holland, 1967).
Wise, George, “Ionists in Industry: Physical Chemistry at General Electric, 1900-1915”,
Isis 74 (1973), 6-21.
Wise, George, “A New Role for Professional Scientists in Industry: Industrial Research at
General Electric, 1900-1916”, Technology and Culture, 21 (1980), 408-429.
Wise, George, Willis R. Whitney, General Electric, and the Origins of Industrial
Research (Columbia University Press, 1985).
Woody, Andrea. I. “Putting Quantum Mechanics to Work in Chemistry: The Power of
Diagrammatic Representation”, Philosophy of Science, 67 (2000) (Proceedings),
S612-S627.
Woody, Andrea, and Glymour, Clark, “Missing Elements: What Philosophers of Science
Might Discover in Chemistry”, in Bhusan, Nalini, and Rosenfeld, Stuart (eds.), Of
Minds and Molecules: New Philosophical Perspectives on Chemistry (Oxford
University Press, 2000), 17-33.